Bidirectional axonal transport of free glycine in identified neurons R3–R14 of Aplysia

The axonal transport of ³H-amino acids was studied in the axons of identified neurons R3–R14 in the parietovisceral ganglion (PVG) of the mollusc Aplysia. The PVG was incubated (3–24 hr) in media containing physiological concentrations of single ³H-amino acids while the isolated nerve was superfused with plain or chemically altered media. The nerve was then sliced into sequential segments for biochemical analyses or fixed for autoradiography. ³H-glycine was transported at 70 mm/day in 6X greater quantities than other amino acids which were transported at <40 mm/day. In the ³H-glycine experiments, >80% of the label transported into the nerve remained as free glycine, comigrating with glycine in thin-layer chromatographs. In autoradiographs of sections 4 mm from the ganglion-nerve barrier, >50% of the silver grains were over R3–R14 axons which occupy <10% of the nerve cross-sectional area. EM autoradiographs confirmed that grains were within R3–R14 and not in surrounding glia. The selective transport of glycine was inhibited by Hg²⁺, by vinblastine and Nocodazole, and by low Ca²⁺ media. Autoradiographs of vinblastine-treated nerves showed a drastic reduction in label over R3–R14 and other axons. Label was also transported retrogradely; this transport rate was similar to the orthograde rate, but 5–10 times less label moved retrogradely. Autoradiographs showed that the retrograde label was localized to R3–R14 axons. This report clearly demonstrates the rapid, selective, and bidirectional transport of a free amino acid and provides further evidence that glycine may be used as a neurochemical messenger by neurons R3–R14.     

Axonal Transport of Choline Lipids in Normal and Regenerating Rat Sciatic Nerve

Abstract: Biochemical methods were used to study the time course of transport of choline phospholipids (labeled by the injection of [³H]choline into the ventral horn of the lumbar spinal cord) in rat sciatic nerve. Autoradiographic methods were used to localize the transported lipid within motor axons. Transported phospholipid, primarily phosphatidylcholine, present in the nerve at 6 h, continued to accumulate over the following 12 days. No discrete waves of transported lipid were observed (a small wave of radioactive phospholipid moving at the high rate would have been missed); the amounts of radioactive lipid increased uniformly along the entire sciatic nerve. In light-microscope autoradiographs, a class of large-caliber axons, presumably motor axons, retained the labeled lipid. Some lipid, even at 6 h, was seen within the myelin sheaths. Later, the labeling of the myelin relative to axon increased. The continued accumulation of choline phospholipids in the axons probably signifies their prolonged release from cell bodies and their retention in various axonal membranes, including the axolemma. The build-up of these phospholipids in myelin probably represents their transfer from the axons to the myelin sheaths surrounding them. When nerves are crushed and allowed to regenerate for 6 or 12 days, choline phospholipids transported during these times enter the regenerating nerve. In light and electron microscope autoradiographs, transported lipid was seen to be localized primarily in the regenerating axons. However, grains overlay the adjacent Schwann cell cytoplasm, indicating transported lipids were transferred from the regenerating axons to the associated Schwann cells. In addition, some cells not associated with growing axons were labeled, suggesting that phosphatidylcholine and possibly acetylcholine, carried to the regenerating axons by axonal transport, were actively metabolized in the terminal, with released choline label being used by other cells. These results demonstrate that axonal transport supplies mature and growing axons and their glial cells with choline phospholipids.     

Signalling organelle for retrograde axonal transport of internalized neurotrophins from the nerve terminal

The retrograde axonal transport of neurotrophins occurs after receptor-mediated endocytosis into vesicles at the nerve terminal. We have been investigating the process of targeting these vesicles for retrograde transport, by examining the transport of [125I]-labelled neurotrophins from the eye to sympathetic and sensory ganglia. With the aid of confocal microscopy, we examined the phenomena further in cultures of dissociated sympathetic ganglia to which rhodamine-labelled nerve growth factor (NGF) was added. We found the label in large vesicles in the growth cone and axons. Light microscopic examination of the sympathetic nerve trunk in vivo also showed the retrogradely transported material to be sporadically located in large structures in the axons. Ultrastructural examination of the sympathetic nerve trunk after the transport of NGF bound to gold particles showed the label to be concentrated in relatively few large organelles that consisted of accumulations of multivesicular bodies. These results suggest that in vivo NGF is transported in specialized organelles that require assembly in the nerve terminal.     

Retrograde Axonal Transport of Locally Synthesized Phosphoinositides in the Rat Sciatic Nerve

Although autoradiography has demonstrated local incorporation of [3H]inositol into axonal phospholipids after intraneural injection, retrograde axonal transport of phosphatidylinositol has only been demonstrated after injection of lipid precursor into the cell body regions (L4 and L5 dorsal root ganglia) of the sciatic nerve. We now report the retrograde axonal transport of inositol phospholipids synthesized locally in the axons. Following microinjection of myo-[3H]inositol into the rat sciatic nerve (50-55 mm distal to L4 and L5 dorsal root ganglia), a time-dependent accumulation of 3H label occurred in the dorsal root ganglia ipsilateral to the injection site. The ratio of dpm present in the ipsilateral dorsal root ganglia to that in the contralateral dorsal root ganglia was not significantly different from unity between 2 and 8 h following isotope injection but increased to 10-12-fold between 24 and 72 h following precursor injection. By 24 h following precursor injection, the ipsilateral/contralateral ratio of the water-soluble label in the dorsal root ganglia still remained approximately 1.0, whereas the corresponding ratio in the chloroform/methanol-soluble fraction was approximately 20. The time course of appearance of labeled lipids in the ipsilateral dorsal root ganglia after injection of precursor into the nerve at various distances from the dorsal root ganglia indicated a transport rate of at least 5 mm/h. Accumulation of label in the dorsal root ganglia could be prevented by intraneural injection of colchicine or ligation of the sciatic nerve between the dorsal root ganglia and the isotope injection site. These results demonstrate that inositol phospholipids synthesized locally in the sciatic nerve are retrogradely transported back to the nerve cell bodies located in the dorsal root ganglia.     

Axonal transport: a quantitative study of retained and transported protein fraction in the cat

The fast axonal transport of proteins was studied in the cat sciatic nerve after injection of [3H]leucine into the spinal ganglion or the ventral horn of the seventh lumbar segment. The amount of transported proteins after ganglion injection was linearly related to the amount of label present at the ganglion. At variable intervals after ganglion or spinal cord injection, the sciatic nerves were sectioned in some experiments. The transport of proteins continued in the peripheral nerve stump in a wavelike manner, but the advancing wave leaves a labeled trail behind. A fraction of this trail corresponds to proteins moving at slower velocities than the velocity of proteins in the wave front. Another fraction of the trail corresponds to molecules retained by the axons. Each nerve segment of 5 mm in length retains 1.5% of the transported proteins, and the profile of retained proteins along the sciatic nerves follows a single exponential function. From the proportion of retained proteins, the concentration of transported proteins at the terminals of branching axons as a function of the branching ratio was estimated. In the case of motor axons innervating the soleus muscle of the cat, the concentration of recently transported proteins at the nerve terminals would be approximately 0.83% of the proteins leaving the spinal cord. This low concentration of transported proteins at the nerve terminals may explain the lability of neuromuscular synapses when axonal transport is decreased or interrupted.     

Position-sensitive detector studies of the axonal transport of a pulse of radioisotope

Axonal transport of a pulse of 35S-methionine-labelled material was studied in vitro in amphibian sciatic nerve using position sensitive detectors. Following formation of a pulse of activity using the cold block technique, the nerve was ligated proximal and distal to the pulse and its movement monitored at room temperature (22.5-23.5 degrees C) for up to 16 h. Material transported in the anterograde direction did so with an average maximum velocity of 147 mm/d. The pulse was found to disperse at an average rate of 0.23 mm/mm travel; however, dispersion was found to vary from preparation to preparation more than would be predicted from experimental error alone. Label was observed to reverse direction at the distal ligature in only 2 of 13 preparations. Reversal of label began within approximately 0.4 h of first arrival, and the most rapidly retrogradely transported material moved at a velocity of 80% that of the most rapidly anterogradely transported material.     

Brain dynein (MAP1C) localizes on both anterogradely and retrogradely transported membranous organelles in vivo

Brain dynein is a microtubule-activated ATPase considered to be a candidate to function as a molecular motor to transport membranous organelles retrogradely in the axon. To determine whether brain dynein really binds to retrogradely transported organelles in vivo and how it is transported to the nerve terminals, we studied the localization of brain dynein in axons after the ligation of peripheral nerves by light and electron microscopic immunocytochemistry using affinity-purified anti-brain dynein antibodies. Different classes of organelles preferentially accumulated at the regions proximal and distal to the ligated part. Interestingly, brain dynein accumulated both at the regions proximal and distal to the ligation sites and localized not only on retrogradely transported membranous organelles but also on anterogradely transported ones. This is the first evidence to show that brain dynein associates with retrogradely transported organelles in vivo and that brain dynein is transported to the nerve terminal by fast flow. This also suggests that there may be some mechanism that activates brain dynein only for retrograde transport.     

Incorporation of axonally transported glycoproteins into axolemma during nerve regeneration

The insertion of axonally transported fucosyl glycoproteins into the axolemma of regenerating nerve sprouts was examined in rat sciatic motor axons at intervals after nerve crush. [(3)H]Fucose was injected into the lumbar ventral horns and the nerves were removed at intervals between 1 and 14 d after labeling. To follow the fate of the “pulse- labeled” glycoproteins, we examined the nerves by correlative radiometric and EM radioautographic approaches. The results showed, first, that rapidly transported [(3)H]fucosyl glycoproteins were inserted into the axolemma of regenerating sprouts as well as parent axons. At 1 d after delivery, in addition to the substantial mobile fraction of radioactivity still undergoing bidirectional transport within the axon, a fraction of label was already associated with the axolemma. Insertion of labeled glycoproteins into the sprout axolemma appeared to occur all along the length of the regenerating sprouts, not just in sprout terminals. Once inserted, labeled glycoproteins did not undergo extensive redistribution, nor did they appear in sprout regions that formed (as a result of continued outgrowth) after their insertion. The amount of radioactivity in the regenerating nerves decreased with time, in part as a result of removal of transported label by retrograde transport. By 7-14 d after labeling, radioautography showed that almost all the remaining radioactivity was associated with axolemma. The regenerating sprouts retained increased amounts of labeled glycoproteins; 7 or 14 d after labeling, the regenerating sprouts had over twice as much of radioactivity as comparable lengths of control nerves or parent axons. One role of fast axonal transport in nerve regeneration is the contribution to the regenerating sprout of glycoproteins inserted into the axolemma; these membrane elements are added both during longitudinal outgrowth and during lateral growth and maturation of the sprout.     

Rapid axoplasmic transport of free leucine

Axoplasmic transport of free 3H-leucine has been studied in vivo in the pike olfactory nerve following application of labeled leucine to the olfactory mucosa. A considerable amount of free 3H-leucine is transported at constant velocity along the axon in the form of a distinct peak. The maximum transport velocity for free 3H-leucine is the same as for rapidly transported 3H-protein (130 and 135 mm/day, respectively, at 19 degrees C). Microtubule inhibitors block or significantly reduce the amount of free 3H-leucine transported, but do not influence the transport velocity. Disruption of the oxygen supply abolishes free 3H-leucine transport, so that this phenomenon cannot be explained by diffusion. The amount of free leucine in the rapidly moving peak decreases with time and distance along the axon and is not detectable after 5 h or more. The transported 3H-leucine is not derived from the circulation or from proteolysis of rapidly transported proteins. This study may help to resolve the controversy over the axoplasmic transport of free amino acids since it shows that free 3H-leucine is transported rapidly but does not travel by rapid axoplasmic transport to the end of axons longer than about 30 mm.     

Depression of fast axonal transport in axons demyelinated by intraneural injection of a neurotoxin fromK. humboldtiana

Tullidinol, a neurotoxin extracted from the Karwinskia humboldtiana fruit, dissolved in peanut oil was injected into the right sciatic nerve of adult cats. The contralateral sciatic nerve received an equivalent volume of peanut oil alone. The fast axonal transport of labeled ([³H]Leucine) protein was studied in sensory and motor axons of both sciatic nerves. The radioactive label was pressure injected either into the L7 dorsal root ganglion or the ventral region of the same spinal cord segment. Several days after the toxin injection, the cat limped and the Achilles tendon reflex was nearly absent in the right hind limb. The amount of transported label was decreased distal to the site of toxin injection. Proximal to this site, the transported material was dammed. Sensory and motor axons showed similar changes. In addition, the toxin produced demyelination and axonal degeneration. Axonal transport and the structure of the axons were normal in the contralateral nerve. Both, Schwann cells and axons of the right sciatic nerve showed globular inclusions, presumably oil droplets containing the toxin. We conclude that Schwann cells and axons as well are tullidinol targets.Departamento de Química. Centro de Investigación y de Estudios Avanzados del IPN.Special issue dedicated to Dr. Sidney Ochs.     

The phosphorylation of kinesin regulates its binding to synaptic vesicles.

Membrane organella are transported bidirectionally in cells, and the axonal transport system has provided an ideal model system for studying this bidirectional transport. Kinesin and cytoplasmic dynein were identified as candidates for the motor molecules of fast axonal transport, which transport organella along microtubules anterogradely and retrogradely. However, the mechanism that controls this bidirectional transport is unknown. Our previous work revealed that kinesin in axons was associated abundantly with anterogradely transported membranous organella, most of which are believed to be precursors of synaptic vesicles and axonal plasma membranes, while the fractions bound to retrogradely transported ones were very small (Hirokawa, N., Sato-Yoshitake, R., Kobayashi, N., Pfister, K. K., Bloom, G. S., and Brady, S. T. (1991) J. Cell Biol. 114, 295-302). Here we demonstrated in vitro that the binding of kinesin to synaptic vesicles was concentration-dependent and saturable and could be released by high salt concentration. When kinesin was phosphorylated by cAMP-dependent protein kinase, its binding to symaptic vesicles was significantly reduced. By motility assay and by statistical analysis using electron microscopy, we further revealed that synaptic vesicles preincubated with phosphorylated kinesin associated less frequently with microtubules than synaptic vesicles preincubated with unphosphorylated kinesin. The phosphorylation of kinesin should therefore play an essential role in regulating the direction of fast axonal transport by inhibiting its binding to membrane organella, thus releasing it from membrane organella at nerve terminals.     

Transport-related amino acid metabolism in germinating barley grains

When eight [¹⁴C]-labelled amino acids were separately injected into the endosperm of germinating (4 days at 20°C) barley (Hordeum vulgare L. cv. Himalaya) grains, the label was rapidly taken up by the scutellum and further transported to the shoot and roots. Some of the amino acids (leucine, lysine and asparagine) were transported in an intact form through the scutellum to the seedling, whilst glutamic acid and aspartic acid were largely converted to glutamine in the scutellum. Proline was mainly transported unchanged, but a small part of the label appeared in glutamine. Arginine was mostly broken down in the scutellum, possibly providing ammonia for the synthesis of glutamine. During further transport in the seedling there was a partial transfer of label from glutamine to asparagine, particularly in the shoot. None of the amino acids used supplied carbon for the synthesis of sucrose, glucose or fructose. Glutamine synthetase activity was particularly high in the scutellum during the period of rapid amino acid transport.     

Posttranslational Protein Modification by Amino Acid Addition in Regenerating Optic Nerves of Goldfish

Previous experiments have demonstrated that 4S RNA, (tRNA), is transported axonally during the reconnection and maturation of regenerating optic nerves of goldfish. The present experiments were performed to determine if tRNA is transported axonally during elongation of these regenerating nerves and whether, as has been demonstrated in other systems, it participates in posttranslational protein modification (PTPM). [3H]Uridine was injected into both eyes of fish with intact optic nerves and 0, 2, 4, or 8 days after bilateral optic nerve cut. Fish were killed 2 days after injection, and [3H]RNA was isolated from retinae and nerves by phenol extraction and ethanol precipitation. [3H]RNA was fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Although the percentage of [3H]4S RNA remained constant in all retinal and control nerve samples, regenerating nerves showed a twofold increase by 6 days after injury, suggesting that [3H]4S RNA is transported axonally in regenerating nerves as early as 6 days after injury. In other experiments, the 150,000-g supernatant of optic nerves was analyzed for incorporation of 3H-amino acids into proteins. No incorporation of 3H-amino acid was found in the soluble supernatant, but when the supernatant was passed through a Sephacryl S-200 column (removing molecules less than 20,000 daltons), [3H]Arg, [3H]Lys, and [3H]Leu were incorporated into proteins. This posttranslational addition of amino acids was greater (1.4-5 times for Lys and 2-13 times for Leu) in regenerating optic nerves than nonregenerating nerves, and the growing tips of regenerating nerves incorporated 5-15 times more [3H]Lys and [3H]Leu into proteins than did the shafts.(ABSTRACT TRUNCATED AT 250 WORDS)     

Components of fast and slow axonal transport in the goldfish optic nerve

The composition of the fast and slow components of axonal transport in the goldfish optic nerve was investigated, using specific radioactive precursors injected into the eye. Tritiated glucosamine and fucose label macromolecules, presumably glycoproteins, which are rapidly transported from the eye to the optic tectum. Material labeled with these precursors is not evident in the slowly transported component. Glucosamine and fucose incorporation are blocked when a protein synthesis inhibitor, acetoxycycloheximide, is injected into the eye concurrently with the precursors. As well as labeling macromolecules, ³H-glucosamine and ³H-N-acetylmannosamine ( a precursor of sialic acids) also label rapidly-transported chloroform-methanol-extractable material which may contain transported glycolipids. Two procedures were used to show that the slow component of axonal transport contains tubulin, a protein characteristic of the microtubules:

• (a) Tracer doses of tritiated colchicine injected into the eye label a wave of radioactivity which moves 0.5 mm/day, the rate of slow axonal transport in the goldfish optic nerve. We believe this wave represents the movement of colchicine which is bound to colchicine-binding protein moving in the slow component of axonal transport.
• (b) Tritiated proline labels a slowly transported protein which is precipitated by vinblastine and has a mobility on polyacrylamide gels comparable to authentic tubulin. These results indicate that the fast and slow components of axonal transport each provide specific chemical substances to the nerve endings.

     

Alteration of ethanolamine glycerophospholipid turnover in trembler dysmyelinating mutant: An analysis of the sciatic nerve by biochemistry and radioautography

The turnover of phospholipids was compared in peripheral nerves of Trembler dysmelinating mutant and control mice, after intraperitoneal and local injection of labeled ethanolamine. In the mutant sciatic nerve, neurochemical analysis showed that [¹⁴C]ethanolamine is incorporated into EGP (ethanolamine glycerophospholipids) of the sciatic nerve at a much higher rate in Trembler mutant than in control mice. Furthermore the decay rate of ¹⁴C-labeled EGP is faster in Trembler than in normal animals. The accelerated turnover of EGP in Trembler sciatic nerve affects the diacyl-EGP while the renewal of the alkenylacyl-EGP (plasmalogens) is slower than in controls. Quantitative radioautographic study at the ultrastructural level corroborate that the initial increase of the label in Trembler nerve fibers was different in axons, Schwann cells and myelin sheaths. EM radioautographs reveal indeed that the high label content observed in Trembler axons takes place preferentially in the myelinated portions of axons and drops within 1 week. In both myelinated and unmyelinated segments of the axons, the majority of the radioactivity was contained in axolemma and smooth axoplasmic reticulum. The 10-fold increase of label found in the myelin sheath of Trembler nerve fibers at 1 day raises the question of the origin of the labeled EGP, either by a stimulated synthesis in Schwann cells or by transfer from axonally transported phospholipids. In contrast, the label of axons, Schwann cells and myelin sheaths of control nerve remains stable during the same period.     

Spinal subarachnoid perfusion in the rat: glycine transport from spinal fluid

Abstract— A method was developed for perfusion of the spinal subarachnoid space in the rat. Bidirectional steady-state fluxes of [¹⁴C]glycine between spinal fluid and plasma were measured. [¹⁴C]glycine clearance from spinal fluid was 5-fold greater than its clearance from plasma. Glycine was transported out of spinal fluid by a saturable process, and the rate of transport was unaffected by the other depressant amino acids, GABA, β-alanine, and taurine. Perfused [¹⁴C]glycine and [³H]GABA distributed in an intracellular compartment in spinal cord. The preparation should be useful for study of the release of these inhibitory amino acids from the intact spinal cord.     

AXONAL TRANSPORT OF SULFATED GLYCOPROTEINS AND MUCOPOLYSACCHARIDES IN THE GARFISH OLFACTORY NERVE

—Application of ³⁵SO₄ to the olfactory mucosa of the long-nosed garfish is found to label sulfated macromolecules which are transported down the olfactory nerve. The transported molecules pass along the nerve as a discrete peak whose leading edge has a transport velocity of 206 ± 6 mm/day. A large portion of the radioactivity from the peak is deposited along the axon. At 2 days after isotope application 83% of the total nerve radioactivity is in the axons and the remaining 17% has accumulated at the terminals in the olfactory bulb. Characterization of sulfated material in the migrating peak indicates that both sulfated glycoproteins (isolated as glycopeptides) and mucopolysaccharides, including chondroitin sulfate and heparan sulfate, are undergoing transport.     

Increased slow transport in axons of regenerating newt limbs after a nerve conditioning lesion

We have previously shown that a nerve conditioning lesion (CL) made 2 weeks prior to amputation results in an earlier onset of limb regeneration in newts. Studies in fish and mammals demonstrate that when a CL precedes a nerve testing lesion, slow component b (SCb) of axonal transport is increased compared to axons that had not received a CL. We wanted to know whether the earlier initiation of limb regeneration after a CL was associated with an increase in SCb transport. The transport of [35S]methionine labeled SCb proteins was measured by using SDS-PAGE, fluorography, and scintillation counting. The rate of transport and quantity of SCb proteins was determined at 7, 14, 21, and 28 days after injection of [35S]methionine into the motor columns of normal; single lesioned (i.e., transection axotomy, amputation axotomy, or sham CL followed by amputation); and double-lesioned limb axons (i.e., nerve transection CL followed 2 weeks later by amputation axotomy). The rate of SCb transport in axons of unamputated newt limbs was 0.19 mm/day. There was an increase in the amount of labeled SCb proteins transported in axons regenerating as the result of a single lesion but no acceleration in the rate of SCb transport, which was 0.21 mm/day in axons that received a sham CL followed by limb amputation. The rate of SCb transport doubled (0.40 mm/day) and the amount of labeled SCb proteins being transported was increased when amputation was preceded by a CL. This study demonstrates that the earlier onset of limb regrowth, seen when amputation follows a CL, is associated with an increased transport of SCb proteins. This suggests that limb regeneration is, in part, regulated by axonal regrowth. We propose that the blastema requires a minimum quantity of innervation before progressing to the next stage of limb regeneration, and that the transport of SCb proteins determines when that quantity will be available.     

Axonal transport of RNA during regeneration of the optic nerves of goldfish.

The distribution of radioactive RNA and RNA precursors in the goldfish optic tecta following intraocular injection of 3H-uridine has been studied during various stages of optic nerve regeneration. 3H-uridine was injected into the posterior chamber of the right eye 17, 30, or 60 days after both optic nerves were crushed. Five were sacrificed at time intervals ranging from 0.5 to 21 days after injection. One day prior to sacrificing, 14C-proline was also injected into the right eye as a marked of fast axonal protein transport. Seventeen to 23 days after crushing, the approximate time of nerve reconnection, the amount of radioactive RNA appearing in the left optic tectum was increased by more than ten times control values. Approximately 30 days after crushing the nerve, when the reconnected nerve is maturing, RNA values were still elevated, but significantly decreased from the earlier stage. By 60 days after crushing the optic nerve, the amounts of RNA in the left tectum was close to normal. Evidence suggesting that, at least, some of the radioactive RNA in the tectum originated from RNA transported along optic axons rather than from RNA synthesized locally in the tectum was provided by autoradiographic experiments. Autoradiograms of paraffin sections taken from the goldfish optic tecta after the intraocular injection of 3H-uridine showed a distribution of grains in a linear pattern, suggesting a distribution over the incoming fibers during the reconnection stage of regeneration. Electron microsocpic autoradiography of glutaraldehyde fixed epoxy sections confirmed that a significant number of grains (shown to be 3H-RNA) were, in fact, over regenerating optic axons. Intracranial injection of 3H-uridine, during the same stage of regeneration, on the other hand, resulted in a distribution of grains, specifically over cell perikaprya. These experiments suggest that during the reconnection phase of nerve regeneration, large amounts of RNA may be carried within regenerating optic axons as they enter the optic tectum.     

The movement of membranous organelles in axons. Electron microscopic identification of anterogradely and retrogradely transported organelles

To identify the structures to be rapidly transported through the axons, we developed a new method to permit local cooling of mouse saphenous nerves in situ without exposing them. By this method, both anterograde and retrograde transport were successfully interrupted, while the structural integrity of the nerves was well preserved. Using radioactive tracers, anterogradely transported proteins were shown to accumulate just proximal to the cooled site, and retrogradely transported proteins just distal to the cooled site. Where the anterogradely transported proteins accumulated, the vesiculotubular membranous structures increased in amount inside both myelinated and unmyelinated axons. Such accumulated membranous structures showed a relatively uniform diameter of 50--80 nm, and some of them seemed to be continuous with the axonal smooth endoplasmic reticulum (SER). Thick sections of nerves selectively stained for the axonal membranous structures revealed that the network of the axonal SER was also packed inside axons proximal to the cooled site. In contrast, large membranous bodies of varying sizes accumulated inside axons just distal to the cooled site, where the retrogradely transported proteins accumulated. These bodies were composed mainly of multivesicular bodies and lamellated membranous structures. When horseradish peroxidase was administered in the distal end of the nerve, membranous bodies showing this activity accumulated, together with unstained membranous bodies. Hence, we are led to propose that, besides mitochondria, the membranous components in the axon can be classified into two systems from the viewpoint of axonal transport: "axonal SER and vesiculotubular structures" in the anterograde direction and "large membranous bodies" in the retrograde direction.