Ostariophysan alarm pheromones: Laboratory and field tests of the functional significance of nitrogen oxides

Article abstract of DOI:10.1023/A:1005445629144

In laboratory and field-trapping studies, we exposed fathead minnows (Pimephales promelas) and finescale dace (Chrosomus neogaeus) to hypoxanthine-3-N-oxide and a suite of structurally and functionally similar compounds in order to determine if: (1) hypoxanthine-3-N-oxide functions as a chemical alarm signal in ostariophysan fishes and (2) the purine skeleton, a structural component, or the nitrogen oxide, a functional component, or both act as the molecular trigger in this chemical alarm signaling system. Minnows and dace exhibited significant antipredator responses when exposed to conspecific skin extract or hypoxanthine-3-N-oxide (increased shoaling and decreased area use) and the functionally similar pyridine-N-oxide (increased shoaling) but not to structurally similar molecules lacking a nitrogen oxide functional group or to a swordtail skin extract control. Field-trapping studies revealed similar results. Traps labeled with fathead minnows skin extract, hypoxanthine-3-N-oxide, or pyridine-N-oxide caught significantly fewer fish than did those labeled with distilled water. These data strongly suggest that the nitrogen oxide functional group acts as the chief molecular trigger in the Ostariophysan alarm pheromone system and that, contrary to previous research, hypoxanthine-3-N-oxide may be one of several possible molecules that function as a chemical alarm signal. Here we report the first example of a single functional group capable of eliciting a suite of behavioral responses.

Full text of DOI:10.1023/A:1005445629144

Journal of Chemical Ecology, Vol. 26, No. 1, 2000
OSTARIOPHYSAN ALARM PHEROMONES:
LABORATORY AND FIELD TESTS OF THE
FUNCTIONAL SIGNIFICANCE OF NITROGEN OXIDES
GRANT E. BROWN,1,* JAMES C. ADRIAN, JR.,2 ERIN SMYTH,1
HEATHER LEET,1 and SCOTT BRENNAN1
1Department of Biological Sciences
2Department of Chemistry, Science and Engineering Center, Union College
Schenectady, New York, 12308
(Received April 7, 1998; accepted August 24, 1999)
Abstract—In laboratory and field-trapping studies, we exposed fathead
minnows (Pimephales promelas) and finescale dace (Chrosomus neogaeus)
to hypoxanthine-3-N-oxide and a suite of structurally and functionally similar
compounds in order to determine if: (1) hypoxanthine-3-N-oxide functions as
a chemical alarm signal in ostariophysan fishes and (2) the purine skeleton,
a structural component, or the nitrogen oxide, a functional component, or
both act as the molecular trigger in this chemical alarm signaling system.
Minnows and dace exhibited significant antipredator responses when exposed
to conspecific skin extract or hypoxanthine-3-N-oxide (increased shoaling and
decreased area use) and the functionally similar pyridine-N-oxide (increased
shoaling) but not to structurally similar molecules lacking a nitrogen oxide
functional group or to a swordtail skin extract control. Field-trapping studies
revealed similar results. Traps labeled with fathead minnows skin extract,
hypoxanthine-3-N-oxide, or pyridine-N-oxide caught significantly fewer fish
than did those labeled with distilled water. These data strongly suggest that
the nitrogen oxide functional group acts as the chief molecular trigger in
the Ostariophysan alarm pheromone system and that, contrary to previous
research, hypoxanthine-3-N-oxide may be one of several possible molecules
that function as a chemical alarm signal. Here we report the first example of
a single functional group capable of eliciting a suite of behavioral responses.
Key Words—Ostariophysan fishes, chemical alarm signals, alarm pheromone,
anti-predator behavior, hypoxanthine-3-N-oxide, nitrogen oxides.
*To whom correspondence should be addressed.
139
0098-0331 00 0100-0139$18.00 0  2000 Plenum Publishing Corporation
/
/
/

140
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
INTRODUCTION
Many fishes of the superorder Ostariophysi (and members of their prey guild)
possess chemical alarm signals or alarm pheromones (Smith, 1992; Mathis and
Smith, 1993a; Brown and Godin, 1997; Brown and Smith, 1997; Chivers and
Smith, 1998). Although controversial (Magurran et al., 1996; Smith, 1997a;
Brown and Godin, 1999b), there exists considerable evidence for the antipredator
function of alarm pheromones for both signal senders and receivers in ostario-
physan fishes (Smith, 1986, 1992, 1997b; Mathis and Smith, 1993b; Brown et
al., 1995a,b; Mathis et al., 1995, 1996; Chivers et al., 1996; Brown and Godin,
1999a). When detected by nearby conspecifics (and sympatric heterospecifics),
the alarm pheromone(s) elicits an alarm response, characterized by an increase
in a variety of antipredator behaviors such as refuging, shoaling, and immobil-
ity (Lawrence and Smith, 1989; Mathis and Smith, 1993b; Brown et al., 1995a;
Brown and Smith, 1997).
Pfeiffer et al. (1985) argued that hypoxanthine-3-N-oxide (H3NO) is the
active component of the ostariophysan alarm pheromone system. Hypoxanthine-
3-N-oxide is a relatively small molecule comprised of a purine skeleton with
a nitrogen oxide functional group at the three position (Figure 1). Pfeiffer
et al. (1985) presented individual black tetras (Gymnocorymbus ternetzi) with
hypoxanthine-3-N-oxide and observed the change in dorsoventral orientation (an
indicator of an alarm response). A significant behavioral response was reported,
which did not differ from that elicited by conspecific skin extract.
There remain a number of unanswered questions regarding the chemical
make-up of the ostariophysan alarm pheromone, however. First, it is unknown
if hypoxanthine-3-N-oxide is the major component of the alarm pheromone or
merely one of a suite of potential chemical signals. Given the diversity of the
superorder Ostariophysi, which includes some 55 families (Moyle and Cech,
FIG. 1. Hypoxanthine-3-N-oxide, with standard purine numbering scheme shown.

OSTARIOPHYSAN ALARM PHEROMONES
141
1988) and almost 72% of all freshwater species (Nelson, 1984), it seems unlikely
that a single compound would be common to all species. Second, hypoxanthine-
3-N-oxide is a highly functionalized molecule, with a variety of functional
groups, providing several potential molecular recognition points. It remains
unknown if a single functional group or combination of functional groups orga-
nized by the purine skeleton acts as the molecular trigger. Finally, Pfeiffer et
al. (1985) used a dorsal–ventral body position as an indicator of an antipreda-
tor response. Magurran et al. (1996) argued that the use of ecologically invalid
measures of an antipredator response could result in false positive results under
laboratory conditions. The use of ecologically valid laboratory and field tests are
required for conclusive results (Magurran et al., 1996; Smith, 1997a; Brown and
Godin, 1999b).
Here, we report the results of a study designed to determine whether the
nitrogen oxide functional group, purine skeleton, or an organized combina-
tion acts as the main signaling agent. Fathead minnows and finescale dace
were exposed to conspecific skin extracts, swordtail skin extract [which lacks
any known ostariophysan alarm pheromone (Pfeiffer, 1977; Mathis and Smith,
1993a], synthesized hypoxanthine-3-N-oxide, and a suite of seven structurally
similar molecules. We also report the first complete spectroscopic characteriza-
tion of hypoxanthine-3-N-oxide.
METHODS AND MATERIALS
Preparation of Hypoxanthine-3-N-oxide and Chemical Characterization
The preparation of hypoxanthine-3-N-oxide (3) was accomplished in two
steps in an overall 55% yield according to the method of Kawashima and
Kumashiro (1969) (Scheme 1). The commercially available 6-methoxypurine (1)
was oxidized exclusively at the three position by the action of hydrogen peroxide
in glacial acetic acid to afford the 6-methoxypurine-3-N-oxide (2). The oxide 2 was
then converted to the desired oxide 3 by alkaline hydrolysis of the methoxy group.
General. Melting points were determined on a Mel-Temp II capillary melt-
ing point apparatus and are uncorrected. Elemental analyses were performed by
Atlantic Microlabs of Norcross, Georgia. Low- and high-resolution mass spec-
SCHEME 1.

142
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
tra were performed by the Mass Spectrometry Service of the University of Illi-
nois. Infrared spectra were recorded on a Mattson Genesis II FTIR spectrometer.
Ultraviolet–visible spectra were recorded on a Hewlett Packard 8452A diode
1
array spectrophotometer. H NMR spectra were determined at 200 MHz; 13C
NMR spectra were determined at 50 MHz and were obtained on a Varian Gem-
ini 200 MHz spectrometer. Chemical shifts are expressed in parts per million (d
units). All starting materials and solvents were purchased from Sigma or Aldrich
and were used as received. In vacuo usually refers to concentration on a Rotava-
por (25 mm Hg, 37–40 ) followed by high vacuum (≤0.1 mm Hg).
6-Methoxypurine-3-N-oxide (2). To a stirred solution of 5.00 g (26.7 mmol)
of 6-methoxypurine (1) in 100 ml of glacial acetic acid was added 50 ml of 30%
aqueous hydrogen peroxide solution, and the reaction mixture was protected from
light. After 14 days the reaction mixture was concentrated in vacuo to afford a yel-
low oil. Triteration of the oil with ≈50 ml of cold 95% ethanol precipitated the oxide
2 as a white powder, which was isolated by filtration, washed twice with ≈50-ml
portions of hot 95% ethanol, and air dried to afford 3.18 g (72%), which was used in
the next step without further purification. An analytical sample was recrystallized
from water and dried in vacuo at 80 C; mp 205–210 C with decomposition [Lit.
216–218 (d) C; (Scheinfeld et al., 1969) the sample explodes if inserted above
220 C]; IR (KBr) 3350, 3142, 2951, 1629, 1488, 1447, 1407, 1323, 1254, 1168,
1
1065, 943, 892, 773, 663, 616 cm ; 200 MHz 1H NMR (D2O, 0.05 M phosphate,
pD 7.20) d 8.54 (s, 1H), 8.26 (s, 1H), 4.17 (s, 3H); 50 MHz 13C NMR (DCl-D2O,
pD 0.09) d 156.1, 150.4, 145.8, 142.7, 129.2, 55.1; HR-MS (EI), m z calcd for
/
C6H6N4O2 (M+) 166.0491, measured 166.0495; UVlHmaxO (pH ≤ 1) 208;lHmaxO (pH
2
2
.
6.76) 226; Anal. calcd for C6H6N4O2 0.5 H2O: C, 41.15; H, 4.03; N, 31.99.
Found: C, 41.11; H, 4.03; N, 31.89.
Hypoxanthine-3-N-oxide (3). A stirred solution of 1.01 g (6.08 mmol) of 2
in 50 ml of 10% NaOH was heated at reflux. After 4 hr, the reaction solution was
cooled with an ice bath and concentrated HCl was added to pH 9.0, decolorized
with charcoal, and filterd through Celite. The solution was then made acidic (pH ≤
1) by the addition of concentrated HCl thus precipitating 3 as a light yellow solid,
which was isolated by filtration. The precipitate was stirred in ≈10 ml of water
and 10% NaOH was added dropwise until the solid was dissolved. The resulting
solution was decolorized with charcoal a second time and filtered through Celite.
The filtrate was cooled in an ice bath and glacial acetic acid was added with gen-
tle stirring until precipitation commenced (pH ≈ 4). The reaction mixture was then
cooled overnight in a refrigerator. The resulting pale yellow crystals were isolated
by filtration and washed successively with anhydrous ethanol and diethyl ether
affording 0.70 g (76%); mp: slow decomposition from 210–300 C [Lit. 210–300
(d) C, (Kawashima and Kumashiro, 1969)]; IR (KBr) 3474, 3120, 3064, 1703,
1
1609, 1454, 1408, 1221, 1099, 903, 797, 684, 624, 560 cm ; 200 MHz 1H NMR
(D2O, 0.05 M phosphate, pD 7.20) d 8.29 (s, 1H), 8.21 (s, 1H); 50 MHz 13C NMR

OSTARIOPHYSAN ALARM PHEROMONES
143
(NaOD-D2O, pD 11.54) d 162.6, 149.2, 147.8, 138.2, 123.9; HR-MS (EI), m z
/
calcd for C5H4N4O2 (M+) 152.0334, measured 152.0335; UV lHmaxO (pH ≤ 1) 210;
max
2
H O
2
.
l
(pH 6.76) 220; Anal. calcd for C5H4N4O2 1H2O: C, 35.30; H, 3.56; N,
32.93. Found: C, 35.61; H, 3.51; N, 32.72.
Experimental Stimuli Preparation
Fathead minnows (Pimephales promelas) were collected from an outlet
pond at SUNY Cobleskill, Cobleskill, New York. Finescale dace (Chrosomus
neogaeus) were collected from Lock 7 Kill, a small tributary stream to the
Mohawk River near Schenectady, New York. Minnows and dace were held in
60-liter aquaria, filled with continuously filtered, dechlorinated tap water (pH
8.0, 18 C) on a 12L : 12D cycle. Swordtails were obtained commercially and
held under similar conditions, except the water temperature was maintained at
approximately 24 C. Fish were fed ad libitum, twice daily with commercial flake
food. Mean ( SE) standard length at time of testing was 5.32 0.18 cm for
minnows and 4.98 0.14 cm for dace.
To generate natural skin extracts, donor fish were killed with a blow to
the head (in accordance with Union College Institutional Animal Care and Use
Protocol #2-27-98). The skin was removed from either side of the donor fish and
immediately placed in 50 ml of chilled glass distilled water. Skin samples were
then homogenized, filtered through glass wool, and the final volumes adjusted
by adding glass distilled water. A total of 27.02 cm2 (in 325 ml), 26.21 cm2
(in 300 ml), and 29.2 cm2 (in 360 ml) was collected for minnows, dace, and
swordtails, respectively. Our final concentrations were similar to those used by
Lawrence and Smith (1989). All skin extracts were frozen in 15-ml aliquots at
20 C until needed. As a control, we also froze glass-distilled water in 15-ml
aliquots at 20 C.
In addition to the hypoxanthine-3-N-oxide, synthesized in the laboratory
(as described above), six other chemicals were obtained from commercial sup-
pliers (Table 1). Hypoxanthine, xanthine, and guanine are structurally similar
purine compounds. The purine ring system contains a fused six-member pyrim-
idine ring and five-membered imidazole ring which share a common side. The
six membered 4(3H)-pyrimidone is almost identically to the pyrimidine moiety
of hypoxanthine-3-N-oxide. The final two compounds, pyridine (PN) and pyri-
dine N-oxide (PNO) are both six-membered heterocycles incorporating a single
nitrogen. We chose these compounds because of their structural or functional
similarities to hypoxanthine-3-N-oxide. It is important to note that hypoxanthine-
3-N-oxide and pyridine N-oxide are related only by the nitrogen oxide functional
group that is absent in the remaining compounds.
We prepared our experimental stimuli by dissolving 0.002 g of the com-
pound of interest in 200 ml of glass-distilled water with stirring for 15 min. The

144
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
TABLE 1. CHEMICAL STRUCTURES OF SYNTHETIC STIMULI USED IN BOTH LABORATORY
AND FIELD TRIALS
Hypoxanthine-3-N-oxide
Pyridine-N-oxide
Guanine
Hypoxanthine
Pyridine
4(3H)-Pyrimidone
Xanthine
resulting solutions were frozen in 15-ml aliquots at 20 C until required. These
concentrations are similar to those used by Pfeiffer et al. (1985) and Mathis and
Smith (unpublished data).
Experimental Protocol
Laboratory Trials. Test tanks consisted of 37-liter aquaria, filled with
dechlorinated tap water (pH 8.0) and containing a gravel substrate and a single

OSTARIOPHYSAN ALARM PHEROMONES
145
airstone, mounted along the back wall of the tank. An additional length of tubing
was attached to the airstone, which allowed us to introduce a chemical stimulus
from behind a black viewing curtain, at a distance of about 3 m (without disturb-
ing the test fish). Test tanks were illuminated, on a 12L : 12D cycle, with 25-W
Sunglow fluorescent lamps. Shoals of four minnows or dace (matched for size)
were placed in the test tanks and allowed a 24-hr acclimation period (N 10
shoals for each treatment condition). Control and experimental trials consisted
of a 10-min prestimulus and a 10 min poststimulus observation period. Prior
to the prestimulus period, 60 ml of tank water was withdrawn from the stimu-
lus injection tube and discarded (to remove any stagnant water in the tube). An
additional 60 ml was withdrawn and retained. At the beginning of the poststim-
ulus period, 5 ml of distilled water (control trials) or 5 ml of conspecific skin
extract (CSE), swordtail skin extract (SWT), or one of the seven synthetic stim-
uli (Table 1) was injected and slowly flushed into the tank using the retained
60 ml of tank water. Dye tests demonstrated that this technique results in the
distribution of chemical stimuli throughout the tank in about 16 sec. Control tri-
als were conducted between 08 : 00 and 11 : 00 hr daily, and experimental trials
were conducted between 13 : 00 and 16 : 00 hr. Control trials were conducted
before experimental trials because any response to the experimental stimuli may
have masked any response to the control stimuli (cf. Lawrence and Smith, 1989;
Hazlett, 1997).
During both the pre- and poststimulus observation periods, we recorded the
position of the fish as “area use” and “shoaling index,” every 15 sec. Area use
was recorded as the position of each fish (1 bottom third of the test tank,
3
top third of the test tank) giving scores ranging from 4 (all fish near the
bottom) to 12 (all fish near the surface). Shoaling index (modified from Mathis
and Smith, 1993b) ranged from 1 (no fish within one body length of each other)
to 4 (all fish within one body length of each other). A decrease in area use and
an increase in shoaling index are indicative of an antipredator response under
laboratory conditions (Mathis and Smith, 1993b; Brown et al., 1995a,b).
For both behavioral measures, we calculated the difference between the pre-
and poststimulus observation periods for the control and experimental trials and
compared them using the Mann-Whitney U test. We also compared the inten-
sity of behavioral responses between the two experimental conditions using the
Mann-Whitney U test.
Field Trials. We conducted a series of field trapping studies at Lincoln’s
Pond and Lake Myocotis, E. N. Huyck Preserve, Rensselaerville, New York,
during July and August 1998, to ensure that the laboratory results were not
simply artifacts (sensu Magurran et al., 1996). For each trapping study, we set
Gee’s Improved minnow traps in trios consisting of one trap labeled with a dis-
tilled water control and two traps labeled with separate experimental stimuli. A
total of eight experimental stimuli were tested in four trapping studies (Table 2).

146
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
TABLE 2. PAIRINGS OF EXPERIMENTAL STIMULI AND DISTILLED WATER CONTROLS
USED IN FIELD TRIALS
Distilled water
Hypoxanthine-3-N-oxide
Fathead minnow skin extract
Distilled water
Pyridine
Pyridine-N-oxide
Distilled water
Xanthine
Guanine
Distilled water
Hypoxanthine
4(3H)-pyrimidone
The pairings of experimental stimuli were randomly assigned. For each study,
a total of 10 trios were set (except the PNO PN+ study, were only nine trios
/
could be set).
We labeled the traps by attaching two cellulose sponges (2 × 2 × 2 cm) to
the funnel end of the trap using stainless steel wire. Sponges were infused with
20 ml of distilled water or the experimental stimulus. Traps were set no less than
15 m apart along the shoreline at a depth of approximately 50 cm, and the order
of stimuli within each trio was randomly assigned. Care was taken to ensure that
each trap within a trio was set on similar substrate. Traps were set for exactly 2
hr, beginning at 8 : 30 AM each day. Wisenden et al. (1995) have demonstrated
that fathead minnows actively avoid areas labeled with conspecific skin extract
for periods of at least 2 hr. This observation suggests that the sponges used in
this study should remain active for at least the 2-hr trapping period. After the
2-hr period, we removed the traps and recorded the number and species of all
fish caught. Because of daily variability in capture rates, each trap study included
a distilled water control and was analyzed independently.
For each study, we compared the number of fish caught per trap between the
distilled water control and the two experimental treatments using the Kruskal-
Wallis test. Individual comparisons were made, if required, using the Mann-
Whitney U test. A x2 test was used to detect any effect of the chemical stimuli
on species composition of fish caught.
RESULTS
Laboratory Trials
We observed significant differences in area use scores (i.e., individuals
remained closer to the substrate) in both minnows and dace when exposed to
conspecific skin extract or H3NO (compared to distilled water controls; Figure
2 and 3; Tables 3 and 4). No significant effects were found with the remaining
experimental stimuli (Figure 2 and 3; Tables 3 and 4). In addition, no significant
difference was observed in the area use response to conspecific skin extract or
H3NO for fathead minnows (Mann-Whitney U, Z
0.15, P 0.88) or in the
area use response of finescale dace to conspecific skin extract, H3NO or PNO
(Kruskal-Wallis, H 0.30, P 0.86).

OSTARIOPHYSAN ALARM PHEROMONES
147
FIG. 2. Mean ( SE) change in area use scores by fathead minnows exposed to distilled
water controls (open bars) and experimental stimuli (solid bars). CSE conspecific skin
extract, H3NO hypoxanthine-3-N-oxide, PNO pyridine-N-oxide, GUAN guanine,
HXAN hypoxanthine, PN+ pyridine, PYR 4(3H)-pyrimidone, XAN xanthine,
*
SWT swordtail skin extract. Significant differences at P < 0.05. Probabilities based
on Mann-Whitney U test. See text and Table 3 for details.
FIG. 3. Mean ( SE) change in area use scores by finescale dace exposed to distilled
water (open bars) and experimental stimuli (solid bars). Abbreviations as in Figure 2.
*Significant differences at P < 0.05. Probabilities based on Mann-Whitney U test. See
text and Table 4 for details.

148
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
TABLE 3. INDIVIDUAL COMPARISONS FOR FATHEAD MINNOW RESPONSE TO
EACH EXPERIMENTAL STIMULUS VERSUS DISTILLED WATER CONTROL BASED ON
MANN-WHITNEY U TESTS
Area use
Shoaling index
a
a
Experimental stimulus
Z
P
Z
P
Fathead minnow skin extract
Hypoxanthine-3-N-oxide
Pyridine-N-oxide
Guanine
Hypoxanthine
Pyridine
4(3H)-Pyrimidone
Xanthine
Swordtail skin extract
2.91
2.21
0.38
1.21
0.34
1.02
0.11
0.72
0.72
<0.05
<0.05
NS
NS
NS
NS
NS
NS
NS
2.31
2.42
3.06
0.64
0.19
0.11
0.38
0.57
0.23
<0.05
<0.05
<0.05
NS
NS
NS
NS
NS
NS
a
Probabilities corrected for multiple comparisons with modified Bonferonni test.
We observed similar tends in shoaling index scores. Both minnows and dace
significantly increased shoal cohesion in response to conspecifics skin extract,
H3NO and PNO (Figures 4 and 5; Tables 3 and 4) but not to the remaining
experimental stimuli (Figures 4 and 5; Tables 3 and 4). Again, there were no
significant differences in the magnitude of the shoaling response between con-
specific skin extract, H3NO or PNO for minnows (H 0.88, P 0.64) or dace
(H 1.94, P 0.38).
TABLE 4. INDIVIDUAL COMPARISONS FOR FINESCALE DACE RESPONSE TO EACH
EXPERIMENTAL STIMULUS VERSUS DISTILLED WATER CONTROL BASED ON
MANN-WHITNEY U TESTS
Area use
Shoaling index
a
a
Experimental stimulus
Z
P
Z
P
Finescale dace skin extract
Hypoxanthine-3-N-oxide
Pyridine-N-oxide
Guanine
Hypoxanthine
Pyridine
4(3H)-Pyrimidone
Xanthine
Swordtail skin extract
2.27
2.19
2.01
0.15
0.11
0.78
0.64
0.11
0.23
<0.05
<0.05
NS
NS
NS
NS
NS
NS
NS
2.91
2.26
2.29
0.19
0.79
0.91
0.04
1.10
0.84
<0.05
<0.05
<0.05
NS
NS
NS
NS
NS
NS
b
a
Probabilities corrected for multiple comparisons with modified Bonferonni test.
Significantly different prior to Bonferonni correction.
b

OSTARIOPHYSAN ALARM PHEROMONES
149
FIG. 4. Mean ( SE) change in shoaling index scores by fathead minnows exposed to
distilled water (open bars) and experimental stimuli (solid bars). Abbreviations as in Fig-
*
ure 2. Significant differences at P < 0.05. Probabilities based on Mann-Whitney U test.
See text and Table 3 for details.
FIG. 5. Mean ( SE) change in shoaling index scores by finescale dace exposed to dis-
tilled water (open bars) and experimental stimuli (hatched bars). Abbreviations as in Fig-
*
ure 2. Significant differences at P < 0.05. Probabilities based on Mann-Whitney U test.
See text and Table 4 for details.

150
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
TABLE 5. OVERALL AND INDIVIDUAL COMPARISONS OF MEAN NUMBER OF FISH CAUGHT
IN TRAPS DURING FIELD TRIALS
Individual
comparisons vs.
Fish caught
(mean SE)
Overall
comparison (H)
distilled water
control traps (Z)
a
b
Trial 1
Distilled water
Fathead minnow
skin extract
Hypoxanthine-3-
N-oxide
10.7 2.2
0.7 0.7
0.0 0.0
H
13.71
Z
2.87
P < 0.002
3.02
P < 0.001
P < 0.001
Z
Trial 2
Distilled water
Pyridine
9.2 2.0
5.4 1.8
H
9.20
Z
P
Z
1.37
0.16
2.78
P < 0.004
Pyridine-N-oxide
0.0 0.0
P < 0.002
Trial 3
Distilled water
Xanthine
4.7 0.7
3.1 1.1
H
P
0.16
0.92
N A
/
Guanine
Trial 4
5.3 1.2
N A
/
Distilled water
Hypoxanthine
6.0 1.7
3.9 0.9
H
P
4.51
0.12
N A
/
4(3H)-pyrimidone
5.1 1.4
N A
/
a
b
Kruskal-Wallis nonparametric ANOVA.
Mann-Whitney U test.
Field Trials
Five species were commonly caught during the field trapping studies: green
sunfish (Lepomis cyanellus), yellow perch (Perca flavescens), emerald shiners
(Notropis atherinoides), finescale dace, and fathead minnows. The results of our
trappings studies confirm those of our laboratory studies. Traps with fathead min-
now skin extract, hypoxanthine-3-N-oxide, and pyridine-N-oxide caught signif-
icantly fewer fish than did their paired distilled water control traps (Table 5).
There were no significant effects of guanine, hypoxanthine, pyridine, 4(3H)-
pyrimidone, or xanthine on the number of fish caught (relative to the paired
distilled water controls; Table 5) or on the distribution of species caught (rel-
ative to the paired distilled water controls; x2 7.01, 7.02, 1.65, 6.89, 3.57,
P > 0.05 for all).

OSTARIOPHYSAN ALARM PHEROMONES
151
DISCUSSION
Our laboratory and field data strongly suggest that the nitrogen oxide func-
tional group is a significant component of the ostariophysan chemical alarm
signal system. In the laboratory studies, we observed significant increases in
antipredator behavior in both fathead minnows and finescale dace when they
were exposed to conspecific skin extract, hypoxanthine-3-N-oxide, and pyridine-
N-oxide, but not to guanine, hypoxanthine, xanthine, 4(3H)-pyrimidone, pyri-
dine, or swordtail skin extract. If the purine skeleton were the critical arrange-
ment, we should have seen an increased antipredator response to any of the
purine compounds (i.e., hypoxanthine, guanine, xanthine), but not to the mono-
cyclic 4(3H)-pyrimidone, pyridine, or pyridine-N-oxide. If a combination of the
purine skeleton and the nitrogen oxide functional group were important, we
would predict a significant antipredator response to hypoxanthine-3-N-oxide and
the conspecific skin extract only. If, however, the presence or absence of the
nitrogen oxide functional group were the molecular trigger, only those com-
pounds containing that functional group (hypoxanthine-3-N-oxide and pyridine-
N-oxide) should elicit a significant behavioral response.
We obtained similar results from the field-trapping experiments, support-
ing our laboratory results. Significantly fewer fish were caught in traps labeled
with fathead minnow skin extract, hypoxanthine-3-N-oxide, and pyridine-N-
oxide compared with the distilled water control traps. Captures in traps labeled
with guanine, hypoxanthine, pyridine, or 4(3H)-pyrimidone were not signifi-
cantly different from those in traps labeled with distilled water. Lincoln’s Pond
and Lake Myocotis contain a diverse prey guild (prey species utilizing simi-
lar microhabitats and subject to the same predation pressures), consisting of
both ostariophysan and nonostariophysan species. The observation that nono-
stariophysan species (sunfish and perch) avoided traps labeled with minnow skin
extract, hypoxanthine-3-N-oxide, and pyridine-N-oxide suggests that they may
also rely on nitrogen oxides as an alarm signaling agent. Yellow perch and green
sunfish have recently been shown to possess analogous chemical alarm signals
(Reehan Mirza, personal communication, personal observations). It is possible
that nitrogen oxides may act as a common signaling agent in a variety of prey
fishes, including nonostariophysans. This hypothesis remains to be tested.
These data also suggest that, contrary to Pfeiffer et al. (1985), hypoxanthine-
3-N-oxide may not be the sole active molecule in the ostariophysan alarm sig-
naling system. Rather, as our data suggest, any compound with a nitrogen-oxide
functional group may act as a potential signaling agent. Pfeiffer et al. (1985)
found no response by black tetras to purported “hypoxanthine-1-N-oxide” (4).
In their studies the analyte samples were prepared by desolusion in dilute HCl
with a final acid concentration of 0.01 M (pH 2). Under these conditions it has
been shown that “hypoxanthine-1-N-oxide” exists as the neutral 1-hydroxyhypo-

152
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
FIG. 6. Acid-Base equilibrium between 1-hydroxyhypoxanthine (5) and hypoxanthine-
1-N-oxide (4).
xanthine (5, Figure 6) with pKa 5.68 0.06 for the hydroxy proton (Parham et
al., 1966). If the final pH of the test sample were never significantly above the
pKa of 5, the test hypoxanthine would remain in the neutral-protonated form,
providing additional, indirect support for our assertion that the nitrogen oxide
functional group serves as the primary molecular trigger.
A comparison of the nitrogen–oxygen bond of pyridine-N-oxide and
1-hydroxyhypoxanthine is also illustrative and further supports our hypothesis.
Beyond the fact that these bonds incorporate the same atoms, nitrogen and oxy-
gen, they are not similar; the nitrogens in both hypoxanthine-3-N-oxide and
pyridine-N-oxide bare a formal positive charge whereas the nitrogen in 1-hy-
droxyhypoxanthine does not. This difference will lead to a significant decrease
in the overall bond polarization between the nitrogen and oxygen. Thus, even the
deprotonated form of 1-hydroxyhypoxanthine would not be predicted to function
as signaling agent. Analytical work is ongoing to investigate this hypothesis.
These data also suggest that the alarm signaling system of ostariophysan
fishes may be affected by acidification of natural waterways. Fathead minnow and
finescale dace held at pH 6.0 do not show a significant response to conspecific skin
extract, but do display a normal antipredator response to the same chemical alarm
signals when held at pH greater than 7.5 (Brown, Adrian, and Lewis, unpublished
data). This may result from one of at least three nonmutually exclusive factors.
Smith and Lawrence (1989) report an overall decrease in activity in fathead min-
nows held at lower pH. A reduced level of activity would make it difficult to detect
a significant alarm response under laboratory conditions. Chronic exposure to acid-
ified waters may result in damage to the olfactory epithelium (Lemly and Smith,
1985, 1987), reducing an individual’s ability to detect chemical signals. Finally,
reduced pH of natural waterbodies may result in a structural change to the chemical
alarm signals, reducing its efficiency as an alarm signal. It is likely that reduced pH
levels (e.g., in acidified waterbodies) result in the protonation of some functional
groups, most notably the nitrogen oxide group.
The acidity constant of the conjugate acid of the N-oxide functionality of 3
(structure 6 in Figure 7) has been measured: pKa 1.2 0.1 (Scheinfeld et al.,
1969); as the pH of the medium decreases, the equilibrium will shift in favor of
the hydroxy form 6 and thus decrease the relative concentration of the depro-

OSTARIOPHYSAN ALARM PHEROMONES
153
FIG. 7. Acid-Base equilibrium between 3-hydroxyhypoxanthine (6) and hypoxanthine-
1-N-oxide (3).
tonated active form of the chemical signal. It would be expected that such a
chemical change would affect this compound’s interaction within the binding
site, thus precipitating a change in the observed behavioral response. Experi-
ments are ongoing to address this question.
Acknowledgments—The authors wish to thank Brian Wisenden and Leo Fleishman for their
comments on earlier versions of the manuscript. Financial support was provided by Union College
to G.E.B. and J.C.A. and by a Merck Summer Scholarship to H.L. The authors wish to dedicate this
manuscript to the memory of Professor R. Jan F. Smith.
REFERENCES
BROWN, G. E., and GODIN, J.-G. J. 1997. Anti-predator response to conspecific and heterospecific
skin extract by threespine sticklebacks: Alarm pheromone revisited. Behaviour 134:1123–1134.
BROWN, G. E., and GODIN, J.-G. J. 1999a. Who dares, learns: Chemical inspection behaviour and
acquired predator recognition in a characin fish. Anim. Behav. 57:475–481.
BROWN, G. E., and GODIN, J.-G. J. 1999b. Chemical alarm signals in wild Trinidadian guppies
(Poecilia reticulata). Can. J. Zool. 77:562–570.
BROWN, G. E., and SMITH, R. J. F. 1997. Conspecific skin extract elicits antipredator behaviour in
juvenile rainbow trout (Oncorhynchus mykiss). Can. J. Zool. 75:1916–1922.
BROWN, G. E., CHIVERS, D. P., and SMITH, R. J. F. 1995a. Localized defecation by pike: A response
to labelling by cyprinid alarm pheromone? Behav. Ecol. 36:105–110.
BROWN, G. E., CHIVERS, D. P., and SMITH, R. J. F. 1995b. Fathead minnows avoid conspecific and
heterospecific alarm pheromone in the faeces of northern pike. J. Fish Biol. 47:387–393.
CHIVERS, D. P., and SMITH, R. J. F. 1998. Chemical alarm signalling in aquatic predator-prey systems:
A review and prospectus. Ecoscience 5:338–352.
CHIVERS, D. P., BROWN, G. E., and SMITH, R. J. F. 1996. The evolution of chemical alarm signals:
Attracting predators benefits alarm signal senders. Am. Nat. 148:649–659.
HAZLETT, B. A. 1997. The organisation of behaviour in hermit crabs: Response to variation in stim-
ulus strength. Behaviour 134:59–70.
KAWASHIMA, H., and KUMASHIRO, I. 1969. Studies of purine N-oxides. III. The synthesis of purine
N-oxides. Bull. Chem. Soc. Jpn. 42:750–755.
LAWRENCE, B., and SMITH, R. J. F. 1989. The behavioral response of solitary fathead minnows,
Pimpephales promelas, to alarm substance. J. Chem. Ecol. 15:209–219.
LEMLY, A. D., and SMITH, R. J. F. 1985. Effects of acute exposure to acidified water on the behavioral
response of fathead minnows, Pimephales promelas, to chemical feeding stimuli. Aquat. Toxicol.
6:25–36.

154
BROWN, ADRIAN, SMYTH, LEET, AND BRENNAN
LEMLY, A. D., and SMITH, R. J. F. 1987. Effects of chronic exposure to acidified water on chemore-
ception of feeding stimuli by fathead minnows (Pimephales promelas): Mechanisms and eco-
logical implications. Environ. Toxicol. Chem. 6:225–238.
MAGURRAN, A. E., IRVING, P. W., and HENDERSON, P. A. 1996. Is there a fish alarm pheromone? A
wild study and critique. Proc. R. Soc. London Ser., B. 263:1551–1556.
MATHIS, A., and SMITH, R. J. F. 1993a. Intraspecific and cross-superorder responses to chemical
alarm signals by brook sticklebacks (Culaea inconstans). Ecology 74:2395–2404.
MATHIS, A., and SMITH, R. J. F. 1993b. Chemical alarm signals increase the survival time of fathead
minnows (Pimephales promelas) during encounters with northern pike (Esox lucius). Behav.
Ecol. 4:260–265.
MATHIS, A., CHIVERS, D. P., and SMITH, R. J. F. 1995. Chemical alarm signals: Predator deterrents
or predator attractants? Am. Nat. 146:994–1005.
MATHIS, A., CHIVERS, D. P., and SMITH, R. J. F. 1996. Cultural transmission of predator recognition
in fishes: Intraspecific and interspecific learning. Anim. Behav. 51:185–201.
MOYLE, P. B., and CECH, J. J. 1988. Fishes: An Introduction to Ichthyology, 2nd ed. Prentice Hall,
Englewood Cliff, New Jersey.
NELSON, J. S. 1984. Fishes of the World, 2nd ed. Wiley-Interscience: New York.
PARHAM, J. C., FISSEKIS, J., and BROWN, G. B. 1966. Purine N-oxides. XVIII. Deamination of ade-
nine N-oxide derivatives. J. Org. Chem. 31:966–968.
PFEIFFER, W. 1977. The distribution of fright reaction and alarm substance cells in fishes. Copeia
1977:653–665.
PFEIFFER, W., RIEGELBAUER, G., MEIER, G., and SCHEIBLER, B. 1985. Effect of hypoxanthine-3-N-
oxide and hypoxanthine-1-N-oxide on central nervous excitation of the black tetra, Gymnoco-
rymbus ternetzi (Characidae, Ostariophysi, Pisces) indicated by dorsal light response. J. Chem.
Ecol. 11:507–523.
SCHEINFELD, I., PARHAM, J. C., MURPHY, S., and BROWN, G. B. 1969. Purine N-oxides. XV. 3-N-
Oxides of adenine and hypoxanthine. J. Org. Chem. 34:2153–2156.
SMITH, R. J. F. 1986. Evolution of alarm signals: role of benefits derived from retaining group mem-
bers or territorial neighbours. Am. Nat. 128:604–610.
SMITH, R. J. F. 1992. Alarm signals in fishes. Rev. Fish Biol. Fish. 2:33–63.
SMITH, R. J. F. 1997a. Does one result trump all others? A response to Magurran, Irving and Hen-
derson. Proc. R. Soc. London, Ser. B. 264:445–450.
SMITH, R. J. F. 1997b. Avoiding and deterring predators, pp. 163–190, in J.-G. J. Godin (ed.).
Behavioural Ecology of Teleost Fishes. Oxford University Press, Oxford.
SMITH, R. J. F., and LAWRENCE, B. 1989. Effects of acute exposure to acidified water on the behav-
ioral response of fathead minnows, Pimephales promelas, to alarm substance (Schreckstoff).
Environ. Toxicol. Chem. 7:329–335.
WISENDEN, B. D., CHIVERS, D. P., BROWN, G. E., and SMITH, R. J. F. 1995. The role of experience in
risk assessment: Avoidance of areas chemically labelled with fathead minnow alarm pheromone
by conspecifics and heterospecifics. Ecoscience 2:116–122.