Laboratory Culture of the Sepiolid Squid Euprymna scolopes: A Model System for Bacteria-Animal Symbiosis

11 Pages·1997·4.8 MB·English

by  R T Hanlon

Reference: Bin/ Bull 192:364-374.(June, 1997)
Laboratory Culture of the Sepiolid Squid
A Model
Euprymna scolopes: System
for Bacteria-Animal Symbiosis
ROGER T. HANLON1,AMNIDCHPAAEULL FV..CDLUANELS:A*P. 3S*USAN E. ASHCRAFT1,
1MarineResourcesCenter, MarineBiologicalLaboratory, WoodsHole, Massachusetts02543:
2MarineScienceCenter, Northeastern University, Nahant, Massachusetts01908;and3Department
ofBiology. H'oodsHoleOceanographicInstitution. H'oodsHole. Massachusetts 02543
Abstract. The small Hawaiian sepiolid Euprymnasco- rium Vibriofischeri has been developed into an experi-
lopes, with its symbiotic luminous bacterium I'ihrio mental marine model in celland molecularbiology(Wei
fischeri. was cultured through one complete life cycle in and Young, 1989; McFall-Ngai and Ruby, 1991; Ruby
4 months. Paralarval squid hatchlings were actively and McFall-Ngai. 1992). Various strains of I'ihrio
planktonic forthe first 20-30 days, afterwhich they set- ftscheri (both natural and mutant) have been cultured
tled and assumed the typical adult mode of nocturnal so that components ofthe symbiotic association can be
activity and diurnal quiescence. Squids were aggressive manipulated experimentally (e.g., Boettcher and Ruby.
predatorsthat preferredactivelyswimmingprey upto2- 1990; Ruby and Asato. 1993: Graftv ill.. 1993; Dunlap
4 timestheirsize;theonlydietthat yieldedgoodsurvival el til.. 1995). However, full development ofthis marine
and rapid growth for paralarvae was large adult mysids. model system has been hampered by the inability tocul-
Survival to settlement was 73% on this diet, whereas it turethe host organism thesquid completely through
was 0%-17% on controls and three other diets. Paralar- its life cycle with some degree ofconsistency and stan-
vae initially lacked both detectable luminescence and I'. dard methodology.
fischericellsin their incipient light organs: all remaining Euprymna scolopes is a very small species that is en-
stages produced luminescence, and their light organs demic to the Hawaiian Islands, where it spends much of
were colonized by apparently pure cultures of >105 I'. its life buried in the sand. There have been no field stud-
tischcri typical ofE. scolopes symbiont strains. Survival iesofitsbehaviorandlifecycle,somostofwhatisknown
from settlementtosexual maturitywas76%. Matingand comesfrom laboratorystudies(Singley, 1983). Ofpartic-
egg layingcommenced at 2 months, yet attempts tocul- ular interest is how these small squids might use their
ture the next laboratory generation of hatchlings were relatively huge light organ in theirdaily lives.
not as successful. The results indicate that the host or- Sincethemid- 980s.severalteamsofresearchershave
1
ganism ofthissymbiosis can soon be cultured with con- studied details ofthe symbiosis by bringing wild-caught
sistencythn '.!"h its brieflifecycle, thusopeningnewav- adults to the laboratory, allowing them to mate and lay
enuesofrest,.: intodevelopmentalaspectsofthissym- eggs, and then using the hatchling squids (i.e.. the para-
biosis. larvae) to explore the infection processaswell as the ini-
Introduction tial development of the light organ. However, thus far
Thesymbioticassociation between the Hawaiian sepi- investigators have been unable to examine events be-
olid Euprymna scolopes and the bioluminescent bacte- yond the first week posthatchingbecause the squid para-
larvae deplete their internal yolk supplies by that time
Received 14November 1446;accepted4 March 1997. and perish; several unreported rearing attempts in the
*Present address: Center for Marine Biotechnology, University of 1990s have failed. Nevertheless, before this model sys-
Maryland BiotechnologyInstitute. Baltimore, Maryland2I2(12. tem was developed, Euprymna had been reared from
364
SQUID CULTURE FOR SYMBIOSIS RESEARCH 365
eggs through a portion ofits life cycle. Euprynina hcrryi setts. Squidswereshipped individually in 3-1 plasticbags
was reared to 2 months in Korea by Choe and Oshima containing 1.5 1 of natural seawater. The seawater was
(1963) and Choe (1966). Arnold ct til. (1972) reared 10 filtered to 10 ^m, heavily aerated, and spiked with 1 g/1
E. scolopes (from a starting number of 26) to 28 days, trisbuffer:the remaining 1.5 1 ofeach bagwas filled with
and two survived to 202 days without reproducing. Sin- oxygen and the bags were fitted in insulated shipping
gley (1983) reared 13 of 16 E. scolopes to 28 days and boxes. The total time spent in these bags by each squid
one to 120 days, although its reported length was very wastypically 21 hours. When thebagsarrived at the lab-
mm
small (<4 mantle length). The encouraging results oratory, thetemperaturewasslightlydepressed (down to
ofArnold etal. ( 1972)and Singley(1983)on E. scolopes, 18-19C), thesalinitywas 31-32 ppt, theconcentration
coupled with the rapid development of this symbiosis of dissolved oxygen was 7-1 1 mg/1, the pH was 8.29-
model system in the last 5 years, led us to initiate the 8.41, and the levels of ammonia and nitrite were low
trialsreported here. Theoverallgoal wastodevelopstan- (however, colorimetric reading were often impossible
dard methods for culturing Euprynina and to learn as- due to interference from ink in the bags). Ammoniawas
pectsofitsbiology thatwouldenhance itssuccessful cul- measured with LaMotte kitsthatarebased on thesalicy-
ture in captivity. We hypothesized that behavioral fea- late method and colorimetric analysis. Nitrite and ni-
tures related to feeding would be most important to the trate were measured with Hach kitsthat also used color-
successful culture ofEuprynina hatchlings through the imetric methods; the precision ofthe ammonia and ni-
paralarval stage, asthey haveprovedtobein manyother trite methods is 0.01 mg/1, and the lowest sensitivity is
squid species (e.g.. Boletzky and Hanlon. 1983; Yang et 0.05 mg/1.
al.. 1986: Hanlon et al.. 1989; Hanlon, 1990; Lee el al.. Adult males were housed individually, and females
1994). individually or in pairs, in small chambers (25 cm
Although we hope to eventually culture successive X 33 cm; water depth 18 cm) and were fed ad libitum
generations of this small, fast-growing cephalopod, a with live shrimp (Palaemonetes sp. or Crangon sp.).
near-term goal is to develop a method for rearing apo- Each chamber had 10 cm ofcrushed coral sand for the
symbiotic hatchlings (i.e.. those deprived of the bacte- squidsto bury in.
rium) through most ofthe life cycle, so that cellularand Matingwasinducedonceperweekbytransferringone
molecular aspects of the complex development of the male into a female's chamber overnight. To avoid dam-
symbiosis can be studied in even greater detail. We re- agingthesquid theyweretransported in small clearglass
port hereourinitial findingsthatEupryninascolopescan vials ofwater rather than in nets. Transferred males al-
be cultured through its brieflife cycle under controlled ways mated, and eggs were generally laid the night after
laboratory conditions, and we highlight some of the the male was removed. In the trials reported here, five
more important needs of this species in captivity. We females and three males produced 13 clutches ofeggs,
also describe our observations on some critical features but we usedonly fiveclutches forthe culturetrials.
ofthe behaviorofthis species, including multiple effects It wasessential thateggsbehandledaslittleaspossible
of light and some aspects of reproductive biology that so that they would develop fully and not hatch prema-
require futureexperimentation. turely. The key was to keep temperature, salinity, pH,
nitrogen levels, and light cyclesassteadyaspossible, and
Materialsand Methods to keep disturbance at a minimum (not bumping tanks,
The term "paralarvae" is used to describe young changing light levels, inspecting eggs, etc.). Adults were
squids from hatching to about 3 weeks ofage; it accu- tkheapttwianteclroqsueda,lirteycwiracsulcaotnisnigstseenatw.aEtegrgssylasitdemosntcooreanlsfuirne-
ormafatnjeul,yve1dn9ii8sl8te)is.ngEauunipdshreaysdnuitlhnteasbs(ecshoelaeovdpieeotsraiioslfsahimanetcYmhobluiennrggsoffatrnhodemfHtaahmra--t gceornstawienerres.lefEtgogsn ltahiedcoonralthaenfdlatmotvanekdwtaollaewrearteedgmenetslhy
scraped from the wall with a glass slide; they generally
inloymSeenpciloaltiudaree ionfVtohses,ord1e97r7)S,epwihoiicdheain(calcucdoersdicnugttlteofitshhe. came offasasingle unit, which wasthen transferredto a
Talhtehotuegrhm "tsrqueuidsq"uiisdcsomarmeonmleymbaepprlsieodftotsheomoersdeeprioTiedus,- amiersshtonceonwtaasinpelra.ceEdggisntwoeeraechnemveesrhexcpoontsaeidnetoraaidrj,aacenndtatno
the eggs so that water was constantly circulating near
thoidea.
them. The eggs were kept in the same water that adults
Broodstockacquisition andeggcare had laid them in.
Adults were collected on Oahu, Hawaii, and shipped Rearingchambersandseawatersystems
viaairfreight to the Marine ResourcesCenterofthe Ma- Theactual rearingchamberforeach paralarval rearing
rine Biological Laboratory in Woods Hole, Massachu- trial was a circular black container, 25 cm in diameter.
366 R. T. HANLON ET AL
helped keepthehatchlingsand prey itemsaway from the
sides. A thin layerofsand coated the mesh bottom.
Two small seawater systems were used in the Marine
Resources Center: one was a completely recirculating
system of 340 1, and the other was an open system of
600 1. Local Woods Hole water was the original source,
and this water was passed through a 1-^m filter and
heated to about 23C. The closed system consisted ofan
A-frame with a shallow tray tank for the rearing experi-
mentsabove(70 cm X 78 cm X 9 cm deep)andadeeper
tank below to house the biological filter. The biological
filter was composed ofcrushed oyster shell with an un-
dergravel filter; the oyster shell was 15 cm deep and
spread over an area of 5250 cnr. The water was then
pumped through a canister with a particulate filter and
then through activated carbon and a UV filter. The flow
ratewasabout 22 1/min, andabout 30% ofthewaterwas
replaced weekly. The open system was constructed sim-
ilarly, with a top tray size of 76 cm X 130 cm X 9 cm
deep and a tank below with 9956 cnr crushed oyster
shell asa substitute substrate. Theopen system wasonly
used forsomesquidsafterDay 49 to help reduce crowd-
ing. Immersion heaters in each system helped maintain
temperature. Water quality was checked two or three
times per week and ammonia, nitrite, and nitrate were
determined accordingtothe methods listed above.
The tanks were situated 3 m from a window that
provided indirect natural light. Overhead fluorescent
fixtures provided indirect light on a 12:12 light cycle.
The typical light level falling on the tray tanks was 2-
8 H'm : MA.
Onesmalltrial (/; = 8 squids)wasperformed inasemi-
closed seawater system at the Marine Science Center of
Northeastern University. Natural seawater was used in
an A-frame that held a total of350 1. Crushed coral was
UV
the biological filter, and no other orcharcoal system
was used. Lighting was by direct overhead incandescent
bulbs (50cm away) on a 12:12 cycle. The paralarvae
Figure1. TheculturetanksforEupryinnti.icolitpe.t.(A)Eachpara- were reared in a black circular PVC chamber (20 cm in
larval rearingchamber(1) held 30 squidsand had a viewingport (2). diameter, 20 cm in height) with 150-^m-mesh sidesand
Waterflowedinthroughtwotubes(3)ineachchamberandexitedver- abare plastic bottom.
ttihcraloluygthhrmoeusghhstchreeseannsd(-4c)ovtoertehdesdireavienbo(5t)t.oTmh,ethwehnitfelorwuelderhoisritzhoentsatlalny- Forbehavioral observations, the top trays were fitted
dard 12 inches. (B)ThecompleteclosedsystemandA-frame,showing withglasspanels forhorizontal viewingintoportionsof
thefilterapparali ihelocationofthebiologicalfiltersubstrate(2) the round rearing chambers. Observations were made
andthechambe: nningadultsandformating(3). from the side and also from the top ofthe tanks, and
night observations were accomplished with a night-vi-
sion device that amplified existing dim light. A very
wbietrhwaas2i0m0m-eMrms-emdesinhasscereaewnatbeortttroaymt(aFnikg.to1)a.cThhieevechsahaml-- wgleoawkorfedrelfilgehcttewdalsigahitmiendthaettrhoeocmei(laipnpgrtooxpirmoatdeulcye0a.3sofwt
low waterdepths ofonly 3-6 cm. Water flowed into the m'1 nA.). A video camera (Sony HiBand 8 mm) was
top through rubber tubing and flowed out through the used to record behavior; the night-vision device could
mesh bottom. Thetwo inflowtubeswerearranged along be fit to the front ofthe camcorderto record nocturnal
the edge to create a gentle circular water flow, which behavior.
SQUID CULTURE FOR SYMBIOSIS RESEARCH 367
Hatchlings. stockingdensities, andfoods spread-plates, which were uniform in appearance for
those animals that produced light, were confirmed to be
andMoisntdievmibduraylosswheartechteradnisnfetrhreedfirwsitt2hhaotfutrhkeedyabrakstceycrlet,o gIa'.nfsisocfheEr.isstcroalionpsetsha(tBocehatrtacchteerriasntidcaRlulybyc,olo1n9i9z0e;lDiughntloarp-
the culture chamber. For individual 30-day feeding tri-
el at.. 1995) by (i) their production ofluminescence, al-
aclhsa,m3b0erh.atTchhleinegxscewpetrieonsstowcekreed oinneeatcrhialroinunwdhirceahri5n0g though at a very low level in culture, (ii) their lumines-
hhaattcchhlliinnggss wweerreetsatkoecnketdo athnedNoanheanttriallabionrawthoircy.h Boynldyay8 clneuonmncilenuermseicsneponoucnsesesottfroatithnheeMJaIu-'.ft2io0si3cn,hdeuraicnedaru(tinoiioi)nntdphurecoierdrue-fcpfirencotgduoscntirntaghi,en
50, squidswereremoved from thesmall roundchambers MJ-215 (Kuo </<//.. 1994; Kuoelal.. 1996).
and thejuveniles were reared in the divided tray tanks,
each ofwhich had dimensionsof35 by 39 cm.
Results
Food items included live zooplankton, crustaceans,
and larval fishes. Mysid shrimps of the genera Mysi- Hatchingsuccess, wateriiuulity, andsystem design
dopsis and Neomysis. and larval fishes, Menidia berry-
Una, were commonly used. The term "postlarval mys- An essential improvement that led to successful cul-
ids" was applied to those that were newly hatched from turewasprovidingaverystableenvironmentsothatem-
the brood sacks of females (Lussier et a/.. 1988). They bryos developed to full term and did not hatch prema-
were ofbody lengths 0.5-1.5 mm, whereasadult mysids turely. Failure to do this affected early trials adversely.
were typically 4-10 mm long. Hatchlings were fed 1- Temperature, light, and pH were nearly constant, and
5 times per day between 0700 and 2300 to ensure ade- the nitrogen levelswere kept within standard limits(i.e.,
quate food availability. Progressively larger shrimp 0.10mg/l for ammonia and nitrite, 20 mg/1 nitrate;
(Crangon and Palaemonetes) were provided tojuvenile Spotte, 1973; Hanlon, 1990). Hatching occurred after
and adult squidsastheygrew. about 18-22 days at 21-23C. Particular care was taken
not to bump the tank or handle the eggs during the last
Luminescencemeasurement andquantification of daysofdevelopment. A samplecheck underadissection
symbioticbacteria microscope indicated that hatchlings had no external
yolk amidst the arms, and normal amounts ofinternal
Lightproducedby hatchlings,juveniles,andsubadults yolk alongthe midgut (Arnold et ai, 1972).
was detected qualitatively with a Turner 20e lumino- Concentrations ofammonia (NH and nitrite (NO
3) :)
meter (Sunnyvale. CA). Individual live animals were remained near the recommended levelsof0.10 mg/1 ex-
placed in 5 ml (50 ml for subadult animal) offilter-ster- cept fortwobriefspikesofammoniato0.40 mg/1 onday
ilized (0.2-^m pore size) natural seawater in 30-ml glass 22and0.20 mg/1 onday 71, andonebriefspikeofnitrite
scintillation vials (300-ml glass beaker for subadult ani- to0.20 mg/1onday46. No mortalitiesorunusualbehav-
mal, placed in a foil-lined funnel to channel light into ior could be correlated with these spikes. Nitrate (NO3)
the luminometer detector), and the light produced was levels were very low throughout the trials, ranging from
recorded for 30 s. Three measurements were taken. The 1 to4 mg/1,wellbelowtherecommendedlevel of20 mg/
data reported (as arbitrary light units, LU, per animal) 1. The pH always stayed above 8.0, and salinity ranged
are the highest ofthe light levels detected for each ani- from 30 to 37 ppt although the mean was near 35 ppt.
mal. Temperature ranged from 21 to 25C, but on one occa-
I'.fischericellscolonizingtheanimal light organ were sion the temperature in the Northeastern University
quantified by plate counts using a seawater complete tank decreased slowly to 4C during the night. The tem-
agar (SWC; Nealson. 1978). To minimize the presence peratureswere returned to 21C in the course of5 h. All
of surface-associated bacteria, hatchlings and juveniles ofthese tropical Hawaiian squidssurvived!
were removed from culture tanks in a minimal volume The utility ofa shallow rearing chamber was evident;
ofseawaterand rinsed by three passages in 5 ml offilter- in particularithelpedconcentratepreyorganismsforthe
sterilized seawater in autoclave-sterilized scintillation very young squids. The system also allowed easy visual
vials. The animals were then homogenized in a 15-ml inspection of the tank, including the behavior of the
Ten Broeck tissue homogenizerwith 1 ml offilter-steril- squidsand theirfood sources.
ized seawater. The homogenate was serially diluted and
plated in quadruplicate on SWC agar plates, which were Behaviorofparalarvaeandjuveniles
incubated at room temperature for 24 h before colonies
were counted. For subadults, the light organ was re- The paralarvae in this study showed two notable be-
moved aseptically and homogenized and handled as havioral features: (i) they were strong swimmers, espe-
above. Representative coloniesofbacteriaarisingon the cially compared to other cephalopod hatchlings ofcom-
368 R. T. HANLON ET AL
parable size; and (ii) they were alternately active and
planktonic as well hie for the first month post-
hatching. The first ! , ;;s validated in field observa-
tions when RTH o a the hatchlings and juveniles
in their nat .n Hawaii in August 1995. Very
smallsqui'* -arttowardsprey20cmawayinonly
a second o a large distance for a paralarval ceph-
alopod.
Diurnal paralarval swimming activity is illustrated in
Figure 2. There were many squidsswimmingduringthe
dayinboththeunsuccessfulandsuccessful feedingtrials.
Nevertheless, even the youngest hatchlings spent a large
proportion oftheir day buried in the sand substrate. By
day 30 in our trials, all ofthe squids remained benthic * J.
during the day like the adults. At all phases of the life
B
cycle, the squids were active throughout the night (only
occasional observations were made from midnight to
0600 h). The behaviorofjuvenileswasseemingly identi-
cal to that ofadults: they remained buried in the sand
duringthedayand hunted forprey at night.
At all life stages. Euprymnu tolerated high densities.
By the time the squids had settled (ca. 30 days) and
grown for a total of40-50 days, their density was quite
high considering that they were still in the small round
rearing chambers. At this time some were transferred to
an adjacent tank system, and in both systemsthe squids
were distributed into rectangular sections measuring 35
X 39 cm with a sand substrate ofabout 1-2 cm. There
werenocasesofcannibalism, which iscommon in many Figure 3. Predation by paralarval squids. (A) A 15-day-old para-
cephalopods. larva subduinga very large mysid. which it ate successfully. (B) A 7-
Unlike the squid in most of the trials, the group of day-old paralarvaeatinga mysidand holdingitinatypical positionat
eight taken to Northeastern University were given no mid-carapace.
sandsubstrate. Thesesquidsneversaton thebottom but
swam all the time. After 2-3 weeks, they settled on the
bare PVC bottom; the five squids that survived in this systemwerebroughttoWoodsHoleand rearedwithoth-
erstoadulthood.
100- Feeding
SuccessfulG1 trialwith
dietofadultmysids Substantial effort was devoted to studying the feeding
f 8 H UnsuccessfulG1 trials behavior of squids, the swimming behavior of various
toOCo) 60- 3D0B -oooo ofvariousdiets pirmepyo,ratnandtthfeinidnitnegrawcatsiotnhsaotfhpartecdhaltionrgsanofdEpirtepy.ryTnhueuim,ovsrt-
lopes are voracious predators that prefer very large prey
40- relativetotheirown size. As indicated in Figure 3, it was
common for hatchlings to successfully attack and ingest
preythatwere much largerthan themselves. Conversely,
QB DD 00 theywerecapableofeatingvery small prey, on theorder
of 0.5 mm. Hatchlings were already strong swimmers
10 20 30 40 and could easily make vigorous forward attacks of 12
Age (days) body lengths in 1 s (as measured from videotape). Field
observations at night in Hawaii also confirmed this
ofFsiqguuirdvsjthai vas.ravcatlivsewliymsmwiinmgmbienhgavidourr,inigndidcaaytliingghtthheopuerrsc.enAtfatgeer strongswimmingability.
30daysall vjuid-, lemainedbenthicduringtheday. To assess different preferences, six diets were tried for
SQUID CULTURE FOR SYMBIOSIS RESEARCH 369
thecritical paralarval period of 1-30 dayspost-hatching. large mysid was held at mid-carapace (Fig. 3B), and it is
Figure 4 indicates the clear results: large adult mysids possible that the squids were biting through the dorsal
were preferred and also produced the best survival; 22 nerve cord to immobilize the mysid. We occasionally
of30 squids survived to settlement. These mysids were saw two squids eating the same large mysid and, more
usually stocked at levels ofabout 2-4 per squid, but the rarely, one squid attacking a mysid while in the process
numbers fluctuated a great deal. Control animals re- of eating another; this occurred only with the smaller
ceived no food and all died by day 8. asdid all 30 squids postlarval mysids. Extremely large mysids (e.g.. more
provided with a smorgasbord oflive zooplankton. Post- than4 times the size ofsquids) were noteaten and prob-
larval mysids, which were expected to be the best food, ablydisrupted thesquidsin small chambers.
produced poor results too; only 5 of50 squids survived Paralarval squids showed considerable interest in fish
to settlement. When postlarval mysids and larval fishes larvae (Menidia) on some days. But the Menidia larvae
were provided together, the mortality was initially high seemed to present problems for capture, and survival of
but stabilized within a week; 5 of30 squids survived to squids on this diet was poor. For example, on day 27,
settlement. Similar results were obtained with irregular squids were presented with fishes that were nearly
mm
mixed diets that consisted of postlarval mysids and 12 long. Thesquidsstrongly pursuedthefish larvae,
fishes, aswell aszooplankton andadult mysids;only4of yet often made 4-6 unsuccessful strikes and sometimes
30 survived to settlement. It is worth emphasizing that as manyas20; recall that squidsofthisageand sizewere
food was presented in abundance in every case, and for well developed and strong swimmers. In a typical feed-
all buta fewofthe 30daysin each trial. ing, only 1-3 squids would have a fish 15 min after 30-
The behavior of squids during active periods (i.e., 40 fishes were added and 30 attacks were observed. By
when in thewatercolumn)seemedto berelated solelyto contrast, if30shrimp(Crangon)were placed in thesame
foraging. Very rarely was a prey item attacked near the tank, commonly 8 squids would have successfully cap-
substrate. Generally the squids preferred to strike up- turedonewithin 5 min. Itwasoften observedthatsquids
ward, which often required that they descend under the could not hold ontoafish even when they madecontact,
prey, then move rapidly towards it at about a 45 degree suggestingthatthesuckerswere notabletograspthe fish
vertical angle. Because their two tentacles were not evi- well. Once captured, fish were harder to subdue than
dent until about 25 dayspost-hatching, squidsused their shrimp; it took on the order of 2-3 min, as compared
eight arms to grasp the prey. Feeding was clearly stimu- with 10-30 s for shrimp. When a fish was being eaten,
lated by the addition ofprey items to the tank; possibly the stomach of the squid was black and highly visible
the increased motion and swimmingactivityofpreywas comparedtoamuch lessdistinctdarkcolorwhenshrimp
enough ofa stimulus to provoke more attack attempts were ingested.
by squids. For this reason it is advisable to feed squids Fooddensities variedgreatly dueto variable food sup-
manytimesperdayratherthanonceorafewtimes. Mys- ply. However, in the early weeks, about 2-4 large adult
ids were clearly preferred, even when other prey such as mysidsweresupplied toeach squid in trial 1. These mys-
fish larvae were present in thetank in fargreaterquanti- ids,which wereabout 1-3 timestheaveragebody length
ties. Squids were commonly observed to make 2-8 of the very young squids, were generally eaten in the
strikes at mysids before a successful capture. Usually a course of24 h. By days 35-40, these squids were being
fed a combination ofadult mysids and Crangon shrimp
(1.0-1.5 cm length) at a level of5-6 times as many prey
as predator; the tanks appeared very crowded when the
food was first put in. By day 52, 15juvenile squidswere
v.-.
being fed about 20 Crangon (1.5cm length) and 100
large mysids.
v ZCoonotprloalnkton Paralarval squids foraged throughout day and night,
x Postlarvalmysids but this was not quantified by counting the number of
V) \A^ Y\ ">-(\^ * AIPHrdoorsusetlogltaualrmravvyraa,sllimdmmisyyxsseiidddssto&&odlIarvalfish abtethaacvkisoronmoprdeeybpyerdauyni3t0,tialmle.sqAufitdesrfsoertatgleindgattonitghhet,adaunldt
\-J\i *-,^\\ ^^ % -----t------- cfloeoadnewdasougtentehrealfloyllaowdidnegdmtoorntianngk.sJluavteeniinletshaenddaaydualntds
continued to feed vigorously on prey that weregenerally
15 20 25 30 theirown size orlarger.
Age (days) Lighting levels appeared to be important to successful
Figure4. Survivalondifferentdietsduringthe I-monthparalarval feeding: lower light enhanced feeding, whereas bright
period. light seemed to retard it. For example, when tops were
370 R. T. HANLON ET AL
placed over the rearing di uiber during the day, the re- weightthroughday 83wasstillbestfitbytheexponential
duced light level stir^ d feeding. Very dark overcast equation y = 0.0033e(""84vl. From this function, the in-
days also resulted quids spending more time stantaneousrelativegrowth ratefrom hatchingtoday83
foraging and f<= iight in the rearing chambers wasan 8.4% increase in weight perday. Afterday 83, the
was arrange adividual squids could view prey data were not amenable to curve fitting because growth
objectsin miglight against a black background, was very slow and thedata were highly variable. Theex-
and this r uiy increased the contrast ofthe prey or- tremely high growth rate ofparalarvae is mostly due to
ganism. the high feeding rate; hatchlings that were feeding on 2-
3verylargemysidsperdaywereprobablyingestingmore
Growth, age, maturity, andmortality than their body weight per day. The length/weight rela-
tionship, based upon 41 measurementsfrom hatchingto
Growth was rapid and adult sizes were reached in day 133, wasexpressed bytheequation
about 2 months. Figure 5 illustrates growth in mantle Wet weight in grams
length and wet weight over the culture period. Hatch-
lings ranged from 1.6 to 1.9 mm in mantle length and = 0.0015 X (mantle length in millimeters)2674
dtfharetoamhisg4eh.tl2y(it.ceoo..n5s.fe8erwvmagptoiivinnetgswrebotewttwwehiemgehenta.sdauDyresesmp1ie0tnetasnthd(ene7la0ir)mlyiatnaelddl pleTtheed cionmpalbeotuetli8fe0cdyacylse,(i.ae.n.dfrtohme elgogngteosetg-gl)ivweadsscqoumi-d
taken from freshly dead squids), rapid growth in wet reached an age of 139 days. As noted in the nextsection,
many squids were sexually mature by 60 days post-
hatching. During the paralarval stage, mortality was
27%, then the population stabilized and the remaining
35
mortalities occurred over a protracted period that was
-y = 0.10215 +021722X R= 091424
30 marked by reproductive activity. From settlingto sexual
maturity, mortalitywasonly24%;thereafter, therewasa
I 25 slow attrition. Early mortalities were probably related to
.c nutrition, but lateroneswere inexplicable and may have
20 -
|o>> been associated with maturity and "old age" or with
15 some unknown pathogen.
1 10
Reproduction
5 First mating & egg laying Sexual dimorphism is only slightlyevident in thisspe-
i cies. Males have slightly enlarged suckers on some arms
20 40 60 80 100 120 140 and tend to have slimmer posterior mantles, especially
Age (days) compared to fully mature femaleswhose posterior man-
tles become broader as they fill with eggs. The testis of
the male can sometimes be seen dorsally when all the
chromatophoresare retracted.
10 The first eggs were laid on day 58 and the first mating
was observed on day 61, when 24 squids were still alive.
S Matingoccurred at night, oftenjust at the onset ofdark-
ness. Overall, 16 matingswereobservedbetweendays61
g>
I and 1 16; the last squid died on day 139, so that mating
-01- occurred over the last one-third to one-halfofthe brief
ID -y = 0.0032968 ' eA(008373x) R= 094481
life cycle. Matings were not controlled in any way and
001 possibly more matings occurred throughout the nights
First mating & egg laying than we observed. Mating (Fig. 6A) seemed to be initi-
ated by the male (bottom), who grasped the female and
0001 placed a spermatophore somewhere in her mantle. Mat-
40 60 80 100 120 140
Age (days) ing lasted about 30-50 min in the fewcases in which the
whole mating was observed. Competition for mates was
ofFmiagumrlee5.lenGgtrhowitnchreoafsceusltvuerresdusEiaigper.yiBtOinTuTsiOWoM/K:'.vW.eTtOPwe:igLhitnegarropwltoht observed only once: two males grabbed a female, some
wasexpi.. '.-ntialthroughday83(R=0.95). Data("ormalesandfemales wrestling followed, then they all moved to cornersofthe
arelumped - RISCidentificationwasnotalwayspossible. tray tank. One male was removed, but the remaining
SOUID CULTURE FOR SYMBIOSIS RESEARCH 371
hatchlings, juveniles, and subadult animals were exam-
ined for the production oflight and for the presence of
r.fischericells. The hatchling squids initially lacked lu-
minescence, and no colony-forming units of I',fischeri
were detected in their nascent light organs. By day 5,
however, the animals were luminous, and their light or-
ganscontained 105ormore I',fischeri'cells(TableI). Fig-
ure 7 illustrates bacterial cells in the light organ of a
reared squid.
Discussion
We have identified three essential keys for successful
laboratory culture of Enprymna scolopes: (i) the eggs
must be provided with conditions that lead to complete
B embryonicdevelopment,thusrenderingfullycompetent
hatchlings; (ii) waterquality must begood, and the tank
configurationand lightingmustbetailoredtothespecific
needs and behavior ofthe species; and (iii) the proper
type and quantity of prey organism must be provided.
Observations in the sea as well as in the laboratory indi-
cated that E. scolopes was a voracious predator for its
size, and that relatively large prey werepreferred.
Becauseour main interest in culturing E. scolopesisto
advance the use of this model ofsymbiosis, it was also
important to demonstrate that laboratory animals were
competent to receive bacteria in a manner similartothat
ofwild-caught Enprymna. Our resultsareconsistent with
previousdemonstrationsthat hatchlings initiallyareapo-
AI/FViYguTrhee6.malMeatgirnagsp(eAd)tahnedfeegmgallaeyifnrgo(mB)unbdyecrunletautrhedaKnidiptrhyeiyunmaa.t<e>d- symbiotic and that they acquire symbiotic strains of I'.
for30-50min. Later,thefemaleaffixedoneeggatatimetoaPVCpipe fischeri, and consequently produce light, within about a
andcoatedtheeggswithsandtocamouflagethem. dayofhatching(Wei and Young, 1989; McFall-Ngai and
Ruby, 1991). Furthermore, squids that had reached the
juvenile benthic and subadult stages also produced lumi-
maleand femaledid not mate that night. In thisparticu- nescence, and their light organs contained increasingly
larcase, the femalewas largerthan the males. largerpopulationsof I'.fischeri(Table I; Fig. 7).
Egg laying (Fig. 6B) was observed four times and, cu-
riously, took place in the morningsand often lasted until
midday. In total, 13 egg clutcheswere laid by thisgener- Table
I
ation ofadults. Attachment ofeach eggtook 10 s on av-
erage, and about 40 s elapsed between the deposition of Re-establishmentofhacicruilsymbiosisinculnircilEuprymna
each egg, so that it took about 25-30 min to lay a clutch scolopes
of30eggs. As usual, theeggsweresoon coated with sand
that wassomehow placed there bythe females. Animalstage
To minimize handling, individual squid were not
marked;thusfecunditycouldbeestimatedonly from the
numberofeggs laid by individual females. All tray tanks
contained males and females in high densities, even
when the squids were separated on day 49; i.e., there
were 2-3 squids per 35 X 39 cm chamber.
Reestablishment ofthebacterial.symbiosis in reared
squids
To determine whether symbiosis with I', fischeri was
reestablished in the light organ of the reared squids.
372 R. T. HANLON ET AL.
tacksofat least 12 body lengths in about a second, a feat
that small squids such asLoligocannot perform.
How might the paralarvaelivein nature?Ourstudy indi-
cates that their activity patterns are flexible compared with
thoseofadults(Fig. 2).Theycanburyinthesandlikeadults,
yet they are surprisingly strong swimmers that can capture
preywithaveryrapidforwardattackandavoidpredatorsby
combiningaswiftbackwardsjetescapewith inking. Wedid
not see any evidence that the light organ was used by para-
larvae,juveniles,oradultsin ourstudy,despitemanyobser-
vationsatnight. Itseemsunlikelythatparalarvaecommonly
use the light organ while seeking prey because we probably
would haveobserved itthroughtheviewingportsintheside
ofthetraytanks(Fig. 1). A moreplausible useforthebiolu-
minescence would be in defense against predators from be-
low,whenEupryinnaofany lifestagearehigherinthewater
column. Since the organ is directed downward, it probably
is used to eliminate or interrupt the squid's shadow against
thedownwellinglight.
Eupryinnascolopeshasoneofthe shortest lifecyclesof
any cephalopod (Boyle, 1983), mainly as a result of its
rapid exponential growth, the warm temperatures it lives
Figure7. (A)Scanningelectron micrograph ofhalfthe lightorgan in, and its small adult size. The fast exponential growth
ofa mature, cultured squid. Letter"p" indicatesthepore. Symbol in-
dicatestheapproximate region ofinterestshown in panel B. (B) High through day 83 istypical for manysquids: since first mat-
magnification of I'ihriulischcri cells within the internal cavity ofthe ingandegglayingwereseen ataboutday 60, weexpected
distalendofthelightorgan. Baris I ^m. that growth would slow substantially soon thereafter,
which it did. The shortest life cycle reported for a cepha-
lopodisinthesmalltropical speciesIdiosepiuspygmaeus:
Behavior, lifecycle, andlaboratoryculturecomparisons statolith ring analysis indicates that this species matures
in 1.5-2.0 monthsand livesonly about 79 days(Jackson,
It hasbeen found repeatedly thatteuthoid and sepioid 1989); this species was not cultured in the laboratory. E.
squids are visual predators that generally preferactively scolopes is relatively easy to rear because of the large
swimmingprey(e.g., Boletzky cta/., 1971; Boletzkyand strong hatchlingsand their propensity for mysid shrimps.
Boletzky, 1973: Boletzky, 1974; Boletzky and Hanlon, Close relatives of E. sco/opes share rapid growth, small
1983; Yang et at.. 1986; Hanlon ct a/.. 1989; Hanlon, size, and preference for mysids. Eupryinna berryi was
1990; Lee el ai. 1994). The peculiar decapod arrange- rearedto2 monthsbyChoeandOshima(1963)andChoe
ment ofeight armsand two tentaclesallowscaptureand (1966); four species ofSepiola and two ofSepiella were
ingestion of both very small and unusually large prey. reared by similar techniques by Boletzky el ai (1971);
Nevertheless, as observed in many of the studies just Sepietla oweniana was cultured by Summers and Berg-
cited, many seemingly good prey itemsare not preferred strom (1981); and Russia macrosoma was reared to
by paralarval squids, thus diets must be determined ex- 8 months by Boletzky and Boletzky (1973). Only E. sco-
perimentally. It was predictable that E. scolopes would lopes, however, lives in warm water, which accelerates
ingest mysidsgiven theenormous success ofthisdiet for growth (Forsythe and Van Heukelem. 1987) and pro-
other squids (studies cited above) and cuttlefish (For- motes a short life cycle. Could these techniques be appli-
sythe el ai, 1994) and the results ofrearing work on E. cableto Eupryinna inorseifrom Japanand tosimilarsep-
herryi(ChoeandOshima, 1963;Choe, 1966)and E. sco- ioidswith light organs?
lopes(Arnold et ai, 1972; Singley, 1983). It was not pre-
dictable that E. scolopes paralarvae would prefer such Behavioralandphysiologicalfactorsthat requirefuture
largepreyand only survivewell on them ascompared to attention
mysids ofothersizes. However, field observations made
with a night-vision device by RTH in Hawaii indicated Nowthat progress has been achieved with the paralar-
thai \ >' young E. scolopes were exceptionally strong val stages, the next logical stage is to focus on reproduc-
swimnK s that couldjet forward at great lengths. Labo- tive behavior. Practically nothing is known about the
ratory \ > in this experiment documented forward at- mating system ofany member ofthe subfamily Sepioli-
SQUID CULTURE FOR SYMBIOSIS RESEARCH 373
naeU'.A'. Boletzky el al.. 1971; Boletzky, 1975; Moyni- biologyofthissymbiosis. The lightorgan requiresbacte-
han, 1983; also reviewed by Hanlon and Messenger. riabeforeitcandevelop(Montgomeryand McFall-Ngai.
1996), and optimal conditions for brood stock manage- 1994), and the bacteria need the light organ to become
mentwill haveto bedetermined before "normal" repro- luminescent (Boettcher and Ruby. 1990). On one hand,
ductive behavior and high fecundity can be expected. the ability to culture the host organism the squid-
The density ofadults, the diet, the light cycle (especially opens the prospect ofstudying late development ofthe
gradual changes that imitate natural changes), the com- light organ, although it first requires that the paralarvae
binations of females and males and the nature oftheir be reared in the absence of the bacteria. On the other
pairings, agonistic behavior, courtship, and any form of hand, onecan nowaddressissuesofinvolvementofbac-
sperm competition will all influence mating, egg laying, teria in developmental programs at the level of tissue,
and the quality ofthe progeny. The matings observed in organ, and whole animal.
this study averaged 35 min; Moynihan (1983) and Sin- The longer-term application of Etiprymna scolopes
gley (1983) reported matings of25-80 min. These long culture is to develop this species into a genetic model.
mating times combined with the presence ofa seminal This is not a trivial task, but E. scolopes has characteris-
receptacle (called the pharetra. located internally near tics that favor success. Unlike other cephalopods. this
the openingofthe oviduct) strongly suggest the possibil- species is small and short-lived, and each female usually
ityofsperm competition behavioramongmales. A prac- lays its clutch ofeggs in a single night and in a discrete
tical problem is to determine how many adult squids clutch,sothatparentageandreproductivesuccesscan be
must be maintained as brood stock under optimal con- assessed.
ditions (physical and social) to ensure sufficient genetic
diversity in subsequent generations. Acknowledgments
A misbalance in the mating system can put stress on We are grateful for advice from John Arnold, Alan
the females, resulting in poor egg production, low fertil- Kuzirian. Bill Mebane. Richard Young, Ned Ruby, and
ization rates, and thus poor hatching rates or lack of Margaret McFall-Ngai. Alan Kuzirian helped produce
vigor in the progeny. Crowded laboratory conditions in the fine SEMs in Figure 7, and John Forsythe helped
cuttlefish and loliginid squids can lead to forced copula- with the growth analysis. Kurt Fiedler collected adults
tions of females (J.G. Boal. pers. comm.. 1996) and a for us in Hawaii. Bill Mebane. Janice Hanley, Louis
disruption of the mating system. Lack of attention to Kerr, and David Remsen ofthe MBL providedessential
these key issues in reproduction is one reason why most logistical support. We thank the Aquatic Resources Di-
cephalopodscultured incaptivityhavenotbeencultured vision of the MBL, Springborn Laboratories (MA),
through multiple generations. Another sepioid. the cut- Aquatic Research Organisms(NH),theMarine Biomed-
tlefish Sepia ojlicinalis. is the one cephalopod that has ical Institute (TX), Dave Bengston (RI), and especially
successfully been cultured through many generations Ray LewisofAquatic Indicators(FL) forhelp in provid-
(Forsythe el at.. 1994), and iscurrently in the 14th labo- ingfoodorganisms. RoxannaSmolowitzperformed nec-
ratory generation at the University of Texas Medical ropsiesofdeadsquids. Thisworkwaspartially fundedby
Branch in Galveston. NSFGrant MCB 94-08266 to PVD.
The second-generation hatchlings in this culture trial
did not do well, but no obvious cause was detected. Be- LiteratureCited
cause we had no access to a control group ofhatchlings
from wild-caught adults, we were unable to determine Arnold. J. M., C. I. Singley, and I.. D. \\illiams-Arnold. 1972.
lwahcekthoefrvitghoerpinrotbhleefmirsltafyilwiailtghenoeurrattieocnh.niWqeueaslsoorhwaidthnoa Es3q6mu1bi-rd3y6Eo4un.pinc'dienvnealouop/mce/n'toaunnddeporslta-bhoartacthoirnygcsounrdviitvialonosf.tIheelsiegpeirol1i4d:
certain cluesto thecauseofthe mortalitiesthat occurred Boettcher, K.J..and E.G. Ruby. 1990. Depressed lightemission by
throughout the trials, although several ofthe adults had symbiotic I'ihriolisclicriofthesepiolid squid Eiiprymnascolopes.
whitish patches ofulceration that are common in many J.Baaeriol. 172:3701-3706.
laboratory-reared cephalopods (Hanlon and Forsythe. Boletzky,S.v. 1974. ElevagedeCephalopodesenaquarium. I'icMi-
//<24(2-A):309-340.
1990). Certainly other factors should be analyzed more Bolelzky,S.v. 1975. ThereproductivecycleofSepiolidae(Mollusca.
closely for example, different light cycles and types of Cephalopoda)./V>W filu: /.out. \iipoli39.Suppl.:84-95.
substrates(cf. Shears. 1988) to see how they affect the Boletzky,S.v.,M.V. \.Boletzky,D.Frosch,andV.Gatzi.1971. Lab-
health and well beingofsquid in captivity. oratory rearingofSepiolinae(Mollusca. Cephalopoda). Mar. Binl
8(1):82-87.
Futurepossibilitieslorthis marinemodel<>lsymbiosis Boleotnzikcy,anSd.ve.a,ralnydpoV.stv-.eBmoblreytoznkiyc.d1e97v3e.lopOmbesnetrvoaftRioosnssioanmtahcene>m\bnrhyi-
The immediate application ofour culture techniques (Mollusca.Cephalopoda).Helgol. ll'i\\. Mccreminiers.S: 135-161.
is in exploring new questions about the developmental Boletzky, S.v., and R. I. Hanlon. 1983. A review ofthe laboratory