Comp. Biochem. Physiol. Vol. 117A, No. 4, pp. 457–462, 1997 ISSN 0300-9629/97/$17.00Copyright 1997 Elsevier Science Inc. PII S0300-9629(96)00398-2

Temperature and Neural Developmentof the Atlantic Herring (Clupea harengus L.)

James Hill and Ian A. JohnstonGatty Marine Laboratory, School of Biological & Medical Sciences,

University of St. Andrews, East Sands, St. Andrews KY16 8LB, Fife, United Kingdom

ABSTRACT. Embryos of Atlantic herring (Clupea harengus L.) from the Buchan (Northern North Sea) stockwere incubated from fertilisation until hatching at temperatures of 5, 8, 12, and 15°C. The relative timing ofdevelopment of the Kolmer-Agduhr (KA) neurons, the posterior lateral line nerve, the motor neurons, andmyotubes were determined with respect to somite stage of the embryo. Development of the KA neurons, thelateral line nerve, and the myotubes was similar at all temperatures. In contrast, timing of outgrowth of themotor neuron axons with respect to somite stage was earlier at higher ($12°C) than at lower temperatures (#8°C) although it reached a similar point at all temperatures by the 58-somite stage. Our hypothesis to explainthese observations is that delayed motor axon outgrowth in the lower temperature groups is probably due to adelay in a signalling interaction between motor neurons and the somite. comp biochem physiol 117A;4:457–462, 1997. 1997 Elsevier Science Inc.

KEY WORDS. Temperature, herring, development, somite, α-acetylated-tubulin, HNK-1, motor neuron, myo-tube, teleost

INTRODUCTION somite formation, construction of myotubes, and the neuraldevelopment of the trunk in embryos reared at a range of

Atlantic herring stocks are distinguishable by their mor-sublethal temperatures. The Buchan stock of Atlantic her-

phology and their spawning grounds, although no signifi-ring spawn at a depth of around 60 m in late August and

cant genetic variation has been observed between thethe eggs develop at a temperature of 11 to 12°C (9). Somite

North Atlantic stocks (12). In addition, each stock spawnsstage was used as a reference to allow us to compare embryos

at a different time of the year, which exposes the embryonicincubated at 5, 8, 12, and 15°C.

stages of each stock to significantly different thermal envi-ronments. Temperature has a major influence upon the de-velopment of ectotherms, affecting the rate and timing of MATERIALS AND METHODSorganogenesis (7). For example, in Spring spawning herring,

Atlantic herring embryos were fertilised and reared as out-the appearance of myofibrillar contractile proteins is rela-lined previously (7,8) and incubated at constant tempera-tively slower in fish reared at 5°C than at 12°C (8). Myofi-tures of 5, 8, 12, and 15°C (60.5°C). A razor blade wasbril assembly is first observed in 27-somite stage embryos atused to scrape eggs from the glass plates at intervals (hours12°C but not until the 42-somite stage at 5°C. The synthe-post fertilisation (hpf)), this was approximately every 2 hrsis of myofibrillar contractile proteins is a direct downstreamat 15°C, and every 2 days at 5°C. The eggs were transferredresult of the expression of the myogenic basic helix-loop-to Petri dishes containing clean sea water and the embryoshelix (bHLH) genes (5) such as MyoD, myogenin and Myf-dissected free from their chorions with fine forceps under a5.dissecting microscope. Embryos older than 30 somites wereSomitogenesis is the earliest morphological evidence ofanaesthetised in 0.02% (w/v) MS222. All embryos weresegmentation in the developing vertebrate (10) and somitefixed in 4% formaldehyde in phosphate buffered salinenumber has been used as a developmental staging tool in(PBS: 0.1 mol l21 phosphate buffer, 0.0027 mol l21 potas-all vertebrates studied. In the present study on North Seasium chloride; 0.137 mol l21 sodium chloride; pH 7.4) over-Atlantic herring (Clupea harengus, L.) from the Buchannight at 4°C. They were then transferred to methanol for(Northern North Sea) stock, we have used antibodieslong term storage at 220°, or to PBS for short-term storage.which label neurons to examine the relationship between

The mouse monoclonal antibody HNK-1 (Sigma, U.K.)recognises a carbohydrate moiety present in both the N-

Address reprint requests to: J. Hill, Gatty Marine Laboratory, University ofCAM and L2 neural cell adhesion molecules (13). It labelsSt. Andrews, East Sands, St. Andrews, KY16 8LB, Fife, United Kingdom.

Tel. 01334 476161; Fax 01334 463443; E-mail: jah1@st-and.ac.uk. a subset of neurons in the developing CNS in zebrafish as

458 J. Hill and I. A. Johnston

FIG. 1. Nomarski interferencecontrast micrographs ofwholemount 50-somite stageherring embryos 8°C stainedwith the HNK-1 antibody.Dorsal is to the top and ante-rior to the right (observationsare consistent in more than 50embryos). (A) Pathway of themotor nerves at mid-trunklevel; arrowheads show themotor roots and the motornerves are clearly seen divid-ing into a dorsal path and aventral path at the lateralchoice points (arrows); (B)Kolmer-Agduhr neurons atthe level of somites 40–45;these cells (arrowheads) de-velop laterally to the floor-plate and extend cilia into theneural canal; (C) Myotubesare clearly visible with No-marski optics, the arrowheadsindicate the nuclei of one my-otube; and (D) The primor-dium of the posterior lateralline nerve at the level of so-mite 30. The arrow shows thepath which the primordiumfollows. Scale bar 50 mm.

they develop axons (14). Anti-α-acetylated tubulin mouse before rinsing 1 3 5 min in PBS and adding primary anti-body. HNK-1 was diluted 1:1000 in GTB; anti-α-ace-monoclonal antibody (Sigma, U.K.) labels processes of all

neurons and has been used to examine the earliest neuronal tylated tubulin was diluted 1:1000 in GTB. Embryos wereincubated overnight at 4°C in primary antibody. The em-pathways in the brain of the zebrafish (2). We stained

whole-mount preparations for tubulin and HNK-1; more bryos were then rinsed 2 hr in GTB (6 changes: 5, 10, 15,and 3 3 30 min) before secondary antibody was added.than 20 embryos from each stage examined were analysed

per temperature group. Anti-mouse IgM HRP-conjugated (HNK-1) and anti-mouse IgG HRP-conjugated (tubulin) were each diluted 1:Embryos were rehydrated from 100% methanol through

graded methanols/PBS. Embryos were stained as whole- 1000 in GTB. The embryos were incubated overnight at4°C and rinsed for 2 hr in PBS (6 changes: 5, 10, 15, andmounts as follows. Non-specific binding sites were blocked

20 min in blocking serum (10% (v/v) normal goat serum 3 3 30 min) before pre-soaking in DAB solution for 20min. H2O2 was added and the stain was allowed to develop.in GTB (1% (w/v) normal goat serum, 1% (v/v) Triton

X100 (Sigma), 1% (w/v) bovine serum albumen in PBS) The staining reaction was stopped by rinsing several times

Temperature and Neural Development 459

FIG. 2. Parasagittal Camera lucida drawing summarising the innervation of the trunk in the herring at the level of somites 4to 7. Anterior is to the right and dorsal is to the top. Somites 5 and 6 have been cut away to reveal the motor nerve anatomy,somites 4 and 7 have been shaded on their superficial surface, and the posterior vertical myosepta of somites 4 and 7 arestippled. The deep dorsal nerve (DDN) and the superficial dorsal nerve (SDN) are the same tract as are the deep ventral nerve(DVN) and the superficial ventral nerve (SVN) (deep nerves are coloured red and superficial nerves are coloured blue forclarity). Motor axons exit the spinal cord through the ventral motor root (VMR) and track ventrally to a choice point (arrow-head), in a similar way to zebrafish motor axons (15). At the choice point each axon either continues to follow the DVN orturns dorsally to follow the DDN. Both nerves follow the perimeter of the vertical myosept (no primary axons were observedpathfinding through the medial region of the somite). In no case (n . 50 embryos studied) did the superficial nerves crossthe horizontal myosept (indicated by the long dashes), the position of which was marked by the lateral line nerve (not shown).The relationship of the choice point to the horizontal myosept has not yet been established. S: spinal cord; N: notochord;scale bar 100 mm.

in PBS. Embryos were mounted in glycerol mounting me- in the ventral spinal cord and were identified as they ex-tended their axon into the somite. The first motor axondium under glass coverslips supported by silicon grease at

each corner. appears in somite 1 and they are added in a rostro-caudalprogression. The motor nerves of the herring initially de-Preparations were examined using a Leitz DMRB system

microscope fitted with Nomarski differential interference velop as a ventral and a dorsal nerve, closely associated withthe periphery of the vertical myoseptum (see legend to Fig.contrast (DIC) optics. The number of specimens used to

establish the neuroanatomy of the herring was . 20 for each 2). The most posterior motor axon which had extended intoits adjacent somite was used to compare the temperaturestructure examined.

The following criteria were noted for each specimen: so- groups. Motor axons stained with α-acetylated-tubulin andtransiently with HNK-1.mite stage; position of the primordium of the posterior lat-

eral line ganglion; position of the most posterior motor KA neurons, previously described in Xenopus (3,4) andzebrafish (1), are present in the ventral spinal cord at theaxon; the position of the most posterior HNK-1 positive KA

neuron; the position of the most posterior myotube (visible level of the neural canal (Fig. 1B). These cells extend shortcilia into the neural canal and have anteriorly projectingunstained with Nomarski optics). Position was determined

by the number of the somite (counting from anterior to pos- axons. They stained transiently with HNK-1 in a rostro-caudal progression as each cell extended its axon.terior) adjacent to each of these structures.

Myotubes could be easily observed with the aid of DICoptics as they developed in a rostro-caudal progression (Fig.

RESULTS1C). The posterior lateral line nerve develops from the Xthcranial ganglion and extends posteriorly as a primordiumThe neural cell types and other structures used for the analy-

sis are shown in Fig. 1. Motor neurons (Fig. 1A) develop (Fig. 1D). The Xth cranial ganglion, the lateral line nerve

460 J. Hill and I. A. Johnston

FIG. 3. Scattergrams and lin-ear regression plots of the po-sition of the most posteriormyotube (A and B), the mostposterior HNK-1 positive Kol-mer-Agduhr neuron (C andD) and the position of the pri-mordium of the posterior lat-eral line nerve (E and F) aredependent on the somite stageof the embryo at all tempera-tures. There is no significanteffect of temperature on therelative rate of the develop-ment of these structures, byanalysis of co-variance. (5°Copen circles and complete line(n . 33); 8°C closed circlesand dotted line (n . 56); 12°Copen squares and dashed line(n . 36) and 15°C closedsquares and dot/dashed line(n . 25)).

and its primordium all stained with both HNK-1 and α- low temperatures than at high temperatures until at approx-imately the 58-somite stage there was no difference betweenacetylated-tubulin.

We found that there was no difference in the timing of the four temperature groups in position of the most posteriormotor neuron relative to somite stage.development of myotubes (Fig. 3A), the KA neurons (Fig.

3B) and the primordium of the posterior lateral line gan- It is also interesting to note that there was almost a directrelationship between the timing of the development of theglion (Fig. 3C) with respect to somite stage at any of the

four temperatures. myotubes and the KA neurons which remained consistentbetween the four temperature groups (Fig. 5).The most anterior motor neuron axons first leave the spi-

nal cord at around the 30-somite stage at 15 and 12°C (nosignificant difference observed) (Fig. 4). In the 5 and 8°C

DISCUSSIONembryos, the first motor axons do not appear in the periph-ery until after the 38- (no motor axons observed in .30 The results demonstrate that in the herring, there is a re-

markable consistency between the temperature groups inembryos) and 40- (no motor axons observed in .30 em-bryos) somite stages, respectively (Fig. 4). Early motor in- the timing of development of the structures studied, i.e.,

they are highly coordinated. The strong relationship be-nervation of the somites proceeds relatively more rapidly at

Temperature and Neural Development 461

tween the timing of HNK-1 expression in the KA neuronsand the development of the myotubes suggests that rostro-caudal development of these structures is also highly coordi-nated probably by factors released by the notochord. In ver-tebrates, the notochord has been shown to be responsiblefor differentiation of both lateral mesoderm (16) and theneural tube (e.g., 18).

In contrast, motor neuron axonogenesis was found tovary with incubation temperature (Fig. 4). At 5 and 8°C,the first motor axons were first observed to leave the spinalcord at later somite stages than at 12 and 15°C. However,the innervation of somites 1 to 35 was more rapid at lowthan at high temperatures, such that motor innervation ofthe somites had reached a comparable stage at all tempera-tures in 58 somite stage embryos (Fig. 4). The first 35 so-

FIG. 4. Motor neuron axonogenesis with respect to somite mites became innervated over 17.6 somite intervals at 12°Cstage in herring embryos. The 8 and 5°C embryos develop compared with only 5.3 somite intervals at 5°C (see legendmotor neurons at later somite stages than the embryos raised to Fig. 4 for an explanation of somite intervals). Our hy-at higher temperatures. The x-axis is the somite interval (si)

pothesis to explain these observations is that the motor neu-from development of the first somite so that the first 62 siron cell bodies probably develop at similar somite stages atcorrespond to the somite stage of the embryo, (si is the time

each temperature group takes to add one somite; the time all temperatures (based on the level of staining of the ven-taken to add one somite was 0.88 hr at 15°C, 0.94 hr at 12°C, tral spinal cord, however, positive identification of the mo-2.24 hr at 8°C, and 2.94 hr at 5°C) (5°C open circles; 8°C tor neurons could only be made after axonogenesis). How-closed circles; 12°C open squares; and 15°C closed squares;

ever, at low temperatures, the signal(s) (probably from themean and SE (most SE are smaller than the symbol) plottedsomite), which causes the motor neuron cell bodies to sendn . 15 for each point).out axons to the periphery, is delayed compared to hightemperatures. On receipt of the appropriate signal(s) themature cell bodies send out axons to the first 20–30 somitesalmost simultaneously at 5°C. Studies in zebrafish haveshown that muscle synapse formation is dependent on acomplex pattern of signalling between motor neurons andmuscle fibres (17). Studies with Clyde (Scottish WestCoast) herring have shown that myofibrils also appear inthe rostral somites at later stages at 5 than 12°C (8). It islikely that the first target for motor axons are mononucleatemuscle pioneer cells which in zebrafish express the homoe-protein engrailed (11) and that the delay in motor axon out-growth observed at low temperatures may be correlated withthe timing of development of these cells.

This work was supported by the Natural Environment Research Coun-cil and is a contribution from the Inter-Universities Marine ResearchInitiative involving the Dunstaffnage Marine Laboratory (Oban) andthe Universities of Dundee, St. Andrews and Stirling. We thank Dr.Peter Tytler for obtaining the herring gonads.

References1. Bernhardt, R.R.; Patel, C.K.; Wilson, S.W.; Kuwada, J.Y. Ax-

onal trajectories and distribution of gabaergic spinal neuronsin wildtype and mutant zebrafish lacking floor plate cells. J.Comp. Neurol. 326:263–272;1992.

2. Chitnis, A.B.; Kuwada, J.Y. Axogenesis by identified neuronsFIG. 5. The direct relationship between development of themost posterior myotube and the most posterior HNK-1 posi- in the brain of zebrafish. J. Neurosci. 10:1892–1905;1990.

3. Dale, N.; Roberts, A.; Ottersen, O.P.; Storm-Mathisen, J. Thetive Kolmer-Agduhr neuron is not affected by temperature.(5°C open circles (n 5 65); 8°C closed circles (n 5 70); 12°C development of spinal cord neurons and their axonal projec-

tions revealed by GABA immunocytochemistry in frog em-open squares (n 5 63); and 15°C closed squares (n 5 41);r . 0.95). bryos. Proc. R. Soc. Lond. B232:205–215;1987.

462 J. Hill and I. A. Johnston

4. Dale, N.; Roberts, A.; Ottersen, O.P.; Storm-Mathisen, J. The 12. King, D.P.F.; Ferguson, A.; Moffet, I.J.J. Aspects of populationgenetics of herring, Clupea harengus, around the British Islesmorphology and distribution of ‘Kolmer-Agduhr cells,’ a class

of cerebrospinal-fluid-contacting neurons revealed in the frog and the Baltic Sea. Fisheries Res. 6:35–52;1987.13. Kruse, J.; Mailhammer, R.; Wernecke, H.; Faissner, A.; Som-spinal cord by GABA immunocytochemistry. Proc. R. Soc.

Lond. B232:193–203;1987. mer, I.; Goridis, C.; Schachner, M. Neural cell adhesion mole-cules and myelin-associated glycoprotein share a common car-5. Davis, R.L.; Weintraub, H.; Lasser, A.B. Expression of a single

transfected cDNA converts fibroblasts to myoblasts. Cell 51: bohydrate moiety recognised by monoclonal antibodies L2and HNK-1. Nature 311:152–155;1984.987–1000;1987.

6. Hempel, G.; Blaxter, J.H.S. The experimental modification of 14. Metcalfe, W.K.; Myers, P.Z.; Trevarrow, B.; Bass, M.B.; Kim-mel, C.B. Primary neurons that express the L2/HNK-1 carbo-meristic characters in herring (Clupea harengus L.). J. Cons.

Int. Explor. Mer. XXVI:336–346;1961. hydrate during early development in the zebrafish. Develop-ment 110:491–504;1990.7. Johnston, I.A. Temperature influences muscle differentiation

and the relative timing of organogenesis in herring (Clupea 15. Myers, P.Z.; Eisen, J.S.; Westerfield, M. Development and ax-onal outgrowth of identified motoneurons in the zebrafish. J.harengus) larvae. Mar. Biol. 116:363–379;1993.

8. Johnston, I.A.; Vieira, V.L.A.; Abercromby, M. Temperature Neurosci. 6:2278–2289;1986.16. Pownall, M.E.; Strunk, K.E.; Emerson, C.P. Notochord signalsand myogenesis in embryos of the Atlantic herring Clupea har-

engus. J. Exp. Biol. 198:1389–1403;1995. control the transcriptional cascade of myogenic bHLH genesin somites of quail embryos. Development 122:1475–1488;9. Jones, S.R.; Jeffs, T.M. Near-surface sea temperatures in

coastal waters of the North Sea, English Channel and Irish 1996.17. Sepich, D.S.; Ho, R.K.; Westerfield, M. Autonomous expres-Sea, Data Report, No. 24. Ministry of Agriculture, Fisheries

and Food Research; 1991. sion of the nic1 acetylcholine recepter mutation in zebrafishmuscle cells. Dev. Biol. 161:84–90;1994.10. Keynes, R.; Stern, C.D. Mechanisms of vertebrate segmenta-

tion. Development 103:413–429;1988. 18. Yamada, T.; Pfaff, S.L.; Edlund, T.; Jessell, T.M. Control ofcell pattern in the neural-tube—motor-neuron induction by11. Kimmel, C.B.; Schilling, T.F.; Hatta, K. Patterning of body

segments of the zebrafish embryo. Curr. Top. Dev. Biol. 25: diffusible factors from notochord and floor plate. Cell 73:673–686;1993.77–110;1991.