Acute Dissociation of Lamprey Reticulospinal Axons to Enable Recording from the Release Face Membrane of Individual Functional Presynaptic Terminals

Our dissociation protocol is significant by yielding isolated reticulospinal axons devoid of postsynaptic projections Figure 2f, but which nevertheless retain functional presynaptic terminals Figure 4c and Figure 4d. The absence of postsynaptic processes opposing the presynaptic terminal permits direct recording access to the presynaptic release face membrane at single presynaptic terminals, previously not possible in central synapses and successfully achieved in only two calyceal-type presynaptic terminals; calyx-type synapse of the chick ciliary ganglion^7-10 and rat calyces^11,12. Individual presynaptic terminals, which at the electron micrographic level have been shown to present simple, single active zones²¹, can be identified by labeling recycling vesicle clusters with FM 1-43 and targeted for recording Figure 4c and Figure 4d. Ca²⁺ currents can be recorded at the release face membrane of individual presynaptic terminals Figure 4f. The recording of Ca²⁺ currents directly from the release face membrane of individual presynaptic terminals is significant as this is the first such recording at central synapses. The physiological relevance of the preparation is emphasized by the fact that the axons retain function after dissociation, allowing for recording from presynaptic terminals with similar characteristics to the terminals in the reticulospinal axons in the intact spinal cord. In addition, presynaptic Ca²⁺ transients in reticulospinal axons have been characterized in the intact lamprey spinal cord through calcium imaging¹³ providing a reference to validate the results obtained from the acutely dissociated axons.

The key to obtaining isolated reticulospinal axons with functional presynaptic terminals lies in a series of critical sequential steps. Tissue impeding acute dissociation of the giant axons located principally in the dorso-medial region of the spinal cord Figure 1a had to be removed first in a vibrating tissue slicer Figure 2a. This makes it easier to mechanically dissociate the spinal cord using the least possible force. The slicing of the dorsal column requires the complete removal of the dorsal meninx primitiva since any remaining dorsal meninx primitiva would impede the cutting action of the blade. We find that leaving the ventral meninx primitiva on the spinal cord at the time of slicing helps maintain the shape of the spinal cord and assists in better slicing. Sustaining tension in the spinal cord during the slicing process prevents any lateral movement or impinging of the tissue under the blade that would damage the axons. 1 cm long spinal cord pieces provides a good length for slicing as the tissue can be stretched and pinned down maintaining tension while slicing. Another critical factor to monitor is the depth of the slice; too shallow a slice prevents good separation of the spinal cord while too deep a slice damages the reticulospinal axons. A good slice depth is one where the dorsal column is removed visibly leaving the ventral reticulospinal column undamaged. The depth of the slice has to be gauged during the slicing process given there is variation in spinal cord thickness between animals. The removal of the dorsal column is best carried out in the middle section of the spinal cord piece Figure 2b as it allows for the unsliced rostral and caudal ends to serve as handles to grasp the tissue during the dissociation process Figure 2e.

After removal of the dorsal column, treating the sliced spinal cord pieces with a mixture of collagenase (Collagenase Type IA from Clostridium histolyticum) and protease (Protease Type XIV from Streptomyces griseus) enzymes (1 mg/ml in Ringer’s solution) helps digest the connective tissue and facilitates a gentler dissociation. A 30-45 min digestion at RT is ideal for dissociation. Complete enzymatic dissociation of the lamprey spinal cord has been previously carried out with sequential treatments of collagenase (2 mg/ml, 30 min) and protease (2 mg/ml, 45 min) to isolate spinal neurons¹⁷. Sequential treatments versus treatment with a cocktail of protease and collagenase did not yield better results in dissociating reticulospinal axons. We also experimented with different concentrations of the dissociation enzymes. Enzyme concentrations higher than 2 mg/ml or digestion times longer than 45 min resulted in the spinal cord losing its consistency and poor dissociations. The higher enzyme concentrations and prolonged treatment periods might aid in complete dissociation of the spinal cord when yields of smaller neurons are required. Reticulospinal axons, in contrast, have an extended structure and treatments greater than 45 min were counterproductive. Retaining some tension in the spinal cord aided better dissociations.

The next important step is to cut the lateral tracts of the enzyme-treated spinal cord pieces at the midpoint of the sliced region to isolate the final dissociation force to the region encompassing reticulospinal axons. The cuts should extend from the lateral edge of the ventral horn gray matter to the central ventro-medial column of reticulospinal axons leaving them undamaged Figure 2d. Cutting the lateral tracts provides a focal point at which forced separation of the spinal cord will occur during mechanical dissociation and facilitates a smoother separation of the tissue. Sustained tension in the spinal cord is required during the cutting process and can be achieved by stretching and pinning the caudal and rostral ends of the spinal cord piece. This prevents tissue deformation under the scalpel blade and consequent damage to the axons.

The final step in the dissociation is applying mechanical tension and gently stretching the tissue along the rostro-caudal axis Figure 2e using forceps to grip the tissue. This step is performed over poly-lysine coated coverslips to allow adhesion of the separated axons for subsequent recording. In order to achieve greater adhesion for the long extended reticulospinal axons, we have used a high molecular weight (>300,000) poly-D-lysine hydrobromide, which provides greater number of attachment points, to coat the coverslips and petri dish insets. The poly-lysine coated surface provides sufficient adhesion for the spinal cord ends and the acutely isolated reticulospinal axons. Teflon coated forceps are required to grip the spinal cord because it sticks firmly to regular stainless steel forceps impeding dissociations. The tissue must be stretched slowly and gently pulling the axons out of the spinal cord. The best way to get the axons to settle down is to maintain the spinal cord ends along the glass coverslip at all times during the dissociation process by maintaining pressure downward onto the rostral and caudal spinal cord ends with the forceps.

The axons are best classified as partially isolated since some portion of the axon remains inside at minimum one end of the spinal cord end depending on the extent of the dissociation. Isolations can be carried out by mechanically separating the two ends of the spinal cord until the axons completely separate from one end of the spinal cord. In this case, one portion of the axon remains inside the spinal cord end from which it emerges while the remaining portion is isolated out of the spinal cord interior. The terminal open end of the axons seals up by a membrane resealing process depending on the health of the axon. Isolations can also be carried out such that the axons emerge out from the spinal cord, but the caudal and rostral ends of the spinal cord remain connected by means of the isolated reticulospinal axons Figure 2e. At this point, gently releasing the pressure allows the axons under tension to relax and loop out yielding isolated regions of axons devoid of postsynaptic projections. Both dissociation techniques yield functional axons. The physiological temperature for the lamprey is 10 °C and hence the preparation is allowed to recover at 10 °C for 1 hr to allow for the axons to recover from the acute dissociation. For electrophysiological recordings, the dissociation was carried out on a poly-lysine coated coverslip inserted into a custom chamber Figure 3. Lamprey spinal cords have been generally maintained at 10 ± 2 °C by perfusing Ringer’s solution, precooled to 10 °C by passing through a thermoelectric cooling device. Exceptionally low background noise is required for resolving single channel Ca²⁺ currents. Perfusion produced artifacts and hence an alternate cooling system had to be designed to maintain the temperature of the dissociated preparation. Recordings were therefore carried out in a still bath containing Ringer’s solution. A custom recording chamber was designed for single channel recordings. The chamber had a central bath compartment where the spinal cord was dissociated in Ringer’s solution to yield isolated axons. A cooling jacket was designed around the central bath chamber to maintain the temperature of the preparation. The temperature of the bath was maintained by perfusing antifreeze solution, precooled by through a thermoelectric cooling device, through the external cooling jacket by passing nitrogen gas under pressure, bypassing any electromechanical pumping requirement (Protocol section 3.15-3.16).

High K⁺ (30 mM KCl) is used to depolarize the axons (section 4.1), so that FM dye is incorporated into vesicles during synaptic exo-endocytosis. It is difficult to retain insertion of a microelectrode in isolated axons for prolonged periods as the slightest drift in axon position causes the electrode to recede from the axon. This makes it difficult to stimulate the axons for prolonged period using a microelectrode. The large axon diameter (20-50 µm) makes it difficult to stimulate using a tungsten stimulation electrode. Hence high K⁺ was used to depolarize the axons. The presynaptic terminals in the reticulospinal axons have a diameter of 1-2 µm.

Ca²⁺ currents are localized to the release face membrane at presynaptic terminals and it is therefore crucial to accurately target the presynaptic terminal. Hence, a key requirement to achieve these recordings is being able to manipulate the movement of the patch pipette position in µm increments. Ca²⁺ currents at presynaptic release face membranes have been demonstrated in calyceal synapses to be extremely small in amplitude, less than 0.5 pA. Single channel recordings of Ca²⁺ currents hence require extremely low background noise levels. Low thermal noise was achieved with an Axopatch 200B amplifier and a cooled headstage. Furthermore, contaminating sources of noise need to be systematically identified and eliminated. Patch pipettes were fabricated, from aluminosilicate glass to ensure low noise since the glass has extremely low impurities and high resistivity. The pipettes were created such that the pipette tip diameter was larger than the diameter of the presynaptic terminal thereby completely encompassing the presynaptic terminal, the dimensions of which can be clearly observed by FM 1-43 labeling. This ensured recording from the entire release face membrane. The capacitive noise was reduced by coating the pipette with PDMS as close to the tip as possible. The gigaohm seals at presynaptic terminal membranes are not stable for long periods and recordings typically last a maximum of 2 min. This makes the recordings extremely difficult to obtain since the low success rate necessitates a large number of attempts to obtain successful recordings. Recording solutions have to be designed so that all other known currents in the presynaptic terminal membrane such as voltage-gated Na⁺ and K⁺ channels and Ca²⁺-activated K⁺ channels are blocked, allowing recording of Ca²⁺ currents.

There are some limitations to using the poly-lysine adhesive surface with this preparation. One is the inability to move the physical location of the chamber containing the preparation because any vibrations disrupted the preparation. Another limitation of the adhesive surface was uncovered during immunohistochemistry experiments, as the rostral and caudal spinal cord end pieces lose adhesion when incubated in 4% paraformaldehyde. We have tried alternative adhesive coatings such as Cell-Tak and encountered similar issues. In order to solve these problems, we used liquid suture glue to affix the spinal cord handles. The specifics regarding the use of the suture glue have been discussed in a previous JoVE article by Chen and Featherstone²². The glue sets at a slower rate in solutions without divalent ions and we took advantage of that property and carried out the dissociations in Ringer’s solution without Ca²⁺ and Mg²⁺. This provided us with additional time for manipulation of the tissue during the dissociation process. We adopted a similar approach to Chen and Featherstone²² using patch pipettes and mouth suction through a tubing connected to the patch pipette to apply the glue onto the specific target area on the dissociation surface prior to performing the dissociation (Protocol section 5.2). The dissociations were performed as described earlier, with the only difference being that the spinal cord ends were affixed during the dissociation process by gently dragging the handle end over the previously placed glue (Protocol section 5.3). Once the dissociation was completed, the divalent ion free Ringer’s solution was exchanged for Ringer’s solution containing Ca²⁺ and Mg²⁺ ions by perfusion and the preparation was allowed to recover at 10 °C for 20-30 min on a thermoelectric cooling plate before proceeding to subsequent steps of immunohistochemistry.

The key implication of this technique is a precise understanding of the Ca²⁺ requirement for neurotransmitter release, which has not been completely resolved despite decades of research. A transient increase in presynaptic [Ca²⁺] has been established as being crucial to triggering release in all synapses. However, consensus is still lacking on the number of Ca²⁺ channels that contribute to the Ca²⁺ domain gating release. Simultaneous measurement of Ca²⁺ currents and capacitance measurements at the release face membrane of individual presynaptic terminals would yield an unambiguous answer to this question. Furthermore, it would enable elucidation of the timing relationship between presynaptic Ca²⁺ entry and vesicle fusion, enabling a precise measure of synaptic delay. The unrestricted accessibility to the release face membrane at individual presynaptic terminals provides the ability to apply a number of different experimental approaches to study various components of neurotransmitter release process. Examples of such approaches include single vesicle imaging using quantum dots to study synaptic vesicle fusion. Vesicle fusion events can also be directly characterized at presynaptic terminals by measuring membrane capacitance changes that underlie vesicle fusion events.