Proc. Nat. Acad. Sci. USAVol. 70, No. 8, pp. 2462-2466, August 1973

Presynaptic Inhibition at Inhibitory Nerve Terminals. A New Synapsein the Crayfish Stretch Receptor

(Procambarus/electron microscopy)

YASUKO NAKAJIMA*, ANN D. TISDALE*, AND MARYANNA P. HENKARTt,4* Department of Biological Sciences, Purdue University, West Lafayette, Indiana, 47907; and t Section onFunctional Neuroanatomy, Laboratory of Neuropathology and Neuroanatomical Sciences, National Institutes ofNeurological Diseases and Stroke, National Institutes of Health, Bethesda, Maryland, 20014

Communicated by Stephen W. Kuffler, May 22, 1973

ABSTRACT Previous physiological evidence has shownthat the receptor neuron of the slowly adapting stretchreceptor organ of crayfish receives synapses from threeinhibitory axons, while the receptor muscle is innervatedby both excitatory and inhibitory axons. Fine structuralstudies have indicated that after certain preparative pro-cedures synaptic vesicles in the inhibitory terminalson dendrites of the receptor neuron appear small andelongate, while those in the excitatory terminals on thereceptor muscle are round and larger. This study describesa new synapse between two inhibitory nerve endings on thereceptor neuron. One axon, containing small elongatevesicles,forms a presynaptic chemical contact withanothermorphologically similar axon that, itself, presumably re-leases inhibitory transmitter onto the receptor neuron.A second type ofpresynaptic axo-axonic synapse, analogousto one previously described in another crustacean muscle,was also found between presumed inhibitory and ex-citatory nerve terminals on the receptor muscle. Thus,the stretch receptor has a relatively complex organizationwith a morphological basis for two types of presynapticinhibition: one on excitatory terminals and the other oninhibitory terminals.

The simplest synaptic inhibitory mechanism in the nervoussystem is secretion by a presynaptic nerve terminal of a chem-ical that reduces the ability of the postsynaptic cell to respondto excitation by another nerve. There exists, in addition,physiological evidence for an inhibition in which one nerveterminal prevents release of an excitatory transmitterfrom another nerve terminal onto a postsynaptic cell. This typeof interaction, called presynaptic inhibition, has been de-scribed in crustacean neuromuscular junction (1) and invertebrate central nervous system (2, 3).

In their electron microscopical study of the opener muscleof crayfish claws, Atwood and Morin (4) found in the vicin-ity of neuromuscular junctions the apparent morphologicalbasis for presynaptic inhibition by describing a typical chem-ical synapse between the inhibitory and motor nerve terminals.

In this preliminary report, we will describe two types of axo-axonic synapses in the stretch receptor organ of crayfish.One occurs on the motor nerve terminal supplying the receptormuscle; it is similar to that reported by Atwood and Morin (4)and corresponds to their presynaptic inhibitory synapse. The

Abbreviations: SVT, small-vesicle terminal; LVT, large-vesicleterminal.t Present address: Behavioral Biology Branch, National In-stitute of Child Health and Human Development, National In-stitutes of Health, Bethesda, Md., 20014.

other is found between two morphologically similar terminalson the stretch receptor neuron. It is of particular interest sinceit suggests the existence of presynaptic inhibition at an in-hibitory synapse, which has not previously been described inany nervous system.

METHODSThe abdominal stretch receptor organs of crayfish(Procambarus) were exposed in van Harreveld's physiologicalsolution (5), fixed with phosphate-buffered 1.6% glutaralde-hyde (pH 7.4), washed with phosphate buffer, and postfixedwith phosphate buffered 1% osmium tetroxide. In most speci-mens all these solutions were isosmotic with the physiologicalsolution (440 milliosmols). In addition, some specimens werefixed and/or washed with solutions that were made eitherhypertonic (880 and 1320 milliosmols) or hypotonic (145milliosmols) to the physiological solution by alteration of thebuffer concentration. After dehydration, specimens were grad-ually infiltrated with epon. During this stage, the slowly adapt-ing receptor organ, easily distinguishable from the fast one byits location and structural characteristics, was separated undera dissecting microscope and embedded. The specimens werethen cut with a Porter-Blum MT2-B or an LKB ultramicro-tome, and examined with a Philips 300 or an AEI2B electronmicroscope.

RESULTS AND DISCUSSION

The anatomy of crustacean stretch receptor organ and itsinnervation have been described by Alexandrowicz (6) andFlorey and Florey (7) and is shown in the diagram (Fig. 1A).The slowly adapting stretch receptor of crayfish consistsof a narrow muscle extending between segments of the exo-

skeleton in the dorsal region of the abdominal and posteriorthoracic segments. In the receptor muscle are embedded thedendrites of a receptor neuron whose axon extends to thecentral nervous system. The receptor neuron dendrites are

innervated by three efferent fibers of different diameters (Ii,I2, and 13 in Fig. 1A or, in the terminology of Alexandrowicz,thick and thin accessory fibers and fiber x). One or possiblytwo of these (I and I3) give off branches that innervate thereceptor muscle which also receives separate motor fibers.There is now good physiological evidence (8-10) that all

efferent innervation of the receptor neuron is inhibitory withthe different inhibitory fibers having different degrees ofeffectiveness. In addition, all synapses on dendrites of the

2462

Proc. Nat. Acad. Sci. USA 70 (1973)

receptor neuron contain small elongate vesicles (11). On thisbasis, the presence of small elongate vesicles in nerve termi-nals has been used as one criterion for distinguishing inhibi-tory from excitatory synapses in crustacean muscle.

Axo-dendritic and neuromuscular synapses

In agreement with the studies of Uchizono (11), Atwood andMorin (4), and Nadol and de Lorenzo (12) we have found thatthe synaptic terminals on the slowly adapting stretch receptorneuron, when fixed and washed in isosmotic solutions, containelongate synaptic vesicles about 330-370 X in diameter (Figs.2 and 4). (Diameter is defined here as the geometric mean ofthe maximum and minimum diameters of each vesicle.) Wewill hereafter refer to this type of terminal, which prob-ably mediates inhibition of the receptor heuron, as SVT (small-vesicle terminal). These are distinguishable from a second typeof terminal that was seen on the stretch receptor muscle only.When prepared under isosmotic conditions, it contains roundvesicles about 410-450 A in diameter. The second type ofterminal probably forms an excitatory neuromuscular junc-tion (13, 14) and we will call it LVT (large-vesicle terminal).SVTs were also found in synaptic contact with the receptormuscle, and these probably correspond to inhibitory synapseson the receptor muscle (14). Evidence that SVTs and LVTscorrespond to inhibitory and excitatory synapses on anothercrayfish muscle is provided by recent experiments of Atwoodet al. (15) in which they selectively depleted SVTs ofvesicles by stimulation of the inhibitor axon and depletedLVTs of vesicles by stimulation of the excitor axon in claw-opener muscle.

Axo-axonic synapsesAtwood and Morin (4) described axo-axonic synapses in thecrayfish claw opener in which SVTs were presynaptic toLVTs, which in turn made a chemical synapse with themuscle. They considered this to be the morphological corre-late of presynaptic inhibition of the excitatory innervation ofclaw-opener muscle. We have observed an analogous rela-tionship in which SVTs are presynaptic to LVTs making asynapse with the receptor muscle (Fig. 4). The diagram in Fig.1C represents this contact.In addition to this type of axo-axonic synapse, we have ob-

served a second kind which has not hitherto been seen. Inthis contact one SVT is presynaptic to another SVT, whichin turn is presynaptic to dendrites of the sensory receptorneuron, as shown in Figs. 2 and 3. These contacts were foundonly in the dendritic regions of the receptor neuron and arecharacterized as chemical synapses by aggregates of smallelongate synaptic vesicles and dense material next to themembrane of the presynaptic axon (white arrows in Figs. 2 and3) and an accumulation of dense material just inside the mem-brane of the adjacent postsynaptic axon. The postsynapticelement of this pair, in turn, contacts another nerve process(D in Fig. 2) by another typical SVT synapse (black arrows inFigs. 2 and 3). This third element does not contain synapticvesicles but usually contains neurotubules, glycogen granules,and mitochondria and is often clearly continuous with a largedendritic process of the receptor neuron. The diagram in Fig.1B represents this type of axo-axonic contact that was ob-served several times in each of several animals.Not only did the pre- and postsynaptic axons at the SVT-

SVT synapse resemble each other in the type of vesicles they

contained, but, in addition, both were found in two cases to bepresynaptic to a single dendrite of a receptor neuron in thesame thin section. Since inhibitory responses have been posi-tively identified in the receptor neuron, while excitatory synap-tic effects have not been recorded, one can be confident thatboth elements of the SVT-SVT synapse are inhibitory to the

A

C

I I I I IRa&n

l lj

ls

FIG. 1. Schematic drawings of the slowly adapting stretchreceptor. (A) Diagram showing the innervation of the slowlyadapting stretch receptor. The dendrites of the receptor neuronreceive three inhibitoryaxons of different diameters. Ip the nomen-clature of Alexandrowicz (6), the thick (IU) and intermediate-size (12) axons are called the thick- and thin accessory fibers,respectively. The thinnest inhibitory axon (I3) is named fiber x.The receptor muscle receives one or two of these inhibitory axons(the thick accessory fiber and probably fiber x) in addition toother motor axons. (B) Schematic drawing of an axo-axonicsynapse found in the dendritic region of the receptor neuron (areaB in part A). This type of axo-axonic synapse is formed by twosmall-vesicle terminals. The postsynaptic terminal of this synapsemakes synaptic contact with the dendrite of the receptor neuron.(C) Schematic drawing of an axo-axonic synapse found on thereceptor muscle (area C in part A). This type of axo-axonicsynapse is formed by a small-vesicle terminal containing smallelongate vesicles and a large-vesicle terminal containing largeround vesicles. The large-vesicle terminal, the postsynaptic com-ponent of this synapse, makes synaptic contact with the receptormuscle.

Crayfish Stretch !receptor 2463

2464 Physiology: Nakajima et at.

,rs tt - 1 Mm* <.%,, i t.: ..

t-so *

... ;... , ,) f .4

/4. ' s g.. _ ,,.

z

R S

s-

ft

.' ;.' . i;

t N4-.

,rF.

I2

fX

ib Wi Ja,.

1

rSVT

its>} *9X~v is ^.

i.e i*; s

~,,

.:z Is.."CN *

.7..

T." , -11..... ,. S,Q4'

I

x "l,

...A

'o

.,tr 'v.-.s ..: -I

tolk,o %T<\

VjtCA. $S

'F 201felei~liot>ji1111 ioip 4io}o oi i'teIX.ol}-axon l nap ni! uo .sfoimS' l1^iel -I sz>1xixs1 si.cn>> t1 1nL ven>1i>Xet t}ew)1^.irfurngal whl1Iticha1.rhe Ot -ldpi

ve-ji le tirno(>~liSn{l 1ilon 11mall i sfiorall Is SA11 11) icxsicle-1' 1B 11li'.titin~ 1 l11^:Iltnl~ik lcoedl xe .icd't>l) Jen-e h~o he-< X 34,600z

Proc. Nat. Acad. Sci. USA 70 (1973)

.aVV A

I.

.1% .. ".1!.1A-rif .;. -

-I

Iff"P.I

:.i

ta.10 .1,

1.4 .-11117

rV'.

t-.

-1 '.

Proc. Nat. Acad. Sci. USA 70 (1973)

~~~~ff :ii~~~~~~~tiA *Y

A~ ~~ %

We\. ~ ~ C

juncion(blckrrow) wth prces oftercptrmsl RM.X460

receptor neuron. A further indication of the similarity of func-tion between the two synaptic components was obtained byvarying the conditions of fixation. It has been shown repeat-edly (16-18) that the size and shape of synaptic vesicles can bechanged by alteration of fixative components and osmolarity,and that controlled use of these changes might be helpful fordistinguishing different types of synapses. We found in ourpreparation that when specimens were washed with a hypo-tonic solution (145 milliosmols) after fixation, vesicles inSVTs and LVTs both appeared round, although those in theSVTs were smaller. On the other hand, when the specimenswere washed in a hypertonic solution (1320 milliosmols) afterfixation the vesicles in both types of terminals appearedelongate, but still of different sizes. In all cases of the SVT-SVT synapses, the size and shape of the synaptic vesicles wereaffected similarly by the tonicity of the wash used. Thus,both components of the SVT-SVT synapse share anatomicaland certain physiological similarities.Our structural evidence suggests that synapses occur be-

tween inhibitory terminals on the receptor neuron. Physio-logical studies are required to clarify the interaction amonginhibitory axons that are efferent to the receptor neuron,particularly since no physiological examples are known in

which the release of inhibitory transmitter from a terminal isinfluenced by another axon.

We thank Drs. S. W. Kuffler, U. J. McMahan, and G. D.Fischbach for their critical reading of this manuscript. Thiswork was supported in part by PHS Grants NS-10457 and NS-08601.

1. Dudel, J. & Kuffler, S. W. (1961) "Presynaptic inhibitionat the crayfish neuromuscular junction," J. Physiol. 155,543-562.

2. Frank, K. & Fuortes, M. G. F. (1957) "Presynaptic andpostsynaptic inhibition of monosynaptic reflexes," Fed.Proc. 16, 39-40.

3. Eccles, J. C. (1964) The Physiology of Synapses (Springer-Verlag, Berlin, Gottingen, Heidelberg).

4. Atwood, H. L. & Morin, W. A. (1970) "Neuromuscular andaxoaxonal synapses of the crayfish opener muscle," J.Ultrastruct. Res. 32, 351-369.

5. Van Harreveld, A. (1936) "A physiological solution forfresh water crustaceans," Proc. Soc. Exp. Biol. Med. 34,428-432.

6. Alexandrowicz, J. S. (1967) "Receptor organs in thoracicand abdominal muscles of crustacea," Biol. Rev. 42, 288-326.

7. Florey, E. & Florey, E. (1955) "Microanatomy of the ab-dominal stretch receptors of the crayfish (AstacusfluviatiliL.)," J. Gen. Physiol. 39, 69-85.

FIG. 3. Electron micrograph showing the axo-axonic synapse (white arrow) formed between two small-vesicle terminals. Procambarusfixed with isosmotic solution and washed with solution twice hypertonic (880 milliosmols) to the physiological solution. The postsynapticterminal in turn makes synaptic contact with a presumed dendrite (D) of the receptor neuron (black arrows). Both small-vesicle terminalscontain small elongate synaptic vesicles. X58,200.

Crayfish Stretch Receptor 2465

2466 Physiology: Nakajima et al.

8. Kuffler, S. W. & Eyzaguirre, C. (1955) "Synaptic inhibitionin an isolated nerve cell," J. Gen. Physiol. 39, 155-184.

9. Burgen, A. S. V. & Kuffler, S. W. (1957) "Two inhibitoryfibers forming synapses with a single nerve cell in the lob-ster," Science 180, 1490-1491.

10. Jansen, J. K. S., Nj&, A., Ormstad, K. & Wall0e, L. (1971)"On the innervation of the slowly adapting stretch receptorof the crayfish abdomen. An electrophysiological approach,"Acta Physiol. Scand. 81, 273-285.

11. Uchizono, K. (1967) "Inhibitory synapses on the stretchreceptor neurone of the crayfish," Nature 214, 833-834.

12. Nadol, J. B., Jr. & de Lorenzo, A. J. D. (1968) "Observa-tions on the abdominal stretch receptor and the fine struc-ture of associated axo-dendritic synapses and neuromus-cular junctions in Homarus," J. Comp. Neurol. 132, 419-443.

Proc. Nat. Acad. Sci. USA 70 (1973)

13. Kuffler, S. W. (1954) "Mechanisms of activation and motorcontrol of stretch receptors in lobster and crayfish," J.Neurophysiol. 17, 558-574.

14. Kosaka, R. (1969) " Electrophysiological and electron micro-scopic studies on the neuromuscular junction of the crayfishstretch receptors," Jap. J. Physiol. 19, 160-175.

15. Atwood, H. L., Lang, F. & Morin, W. A. (1972) "Synapticvesicles: selective depletion in crayfish excitatory and in-hibitory axons," Science 176, 1353-1355.

16. Bodian, D. (1970) "An electron microscopic characteriza-tion of classes of synaptic vesicles by means of controlledaldehyde fixation," J. Cell Biol. 44, 115-124.

17. Nakajima, Y. (1971) "Fine structure of the medial nucleusof the trapezoid body of the bat with special reference totwo types of synaptic endings," J. Cell Biol. 50, 121-134.

18. Valdivia, 0. (1971) "Methods of fixation and the morphologyof synaptic vesicles," J. Comp. Neurol. 142, 257-274.