Here, we described a protocol to quantitatively study the assembly and structure of the axon initial segments (AIS) of hippocampal neurons that lack pre-assembled AIS due to the absence of a giant ankyrin-G.
Neuronal axon initial segments (AIS) are sites of initiation of action potentials and have been extensively studied for their molecular structure, assembly and activity-dependent plasticity. Giant ankyrin-G, the master organizer of AIS, directly associates with membrane-spanning voltage gated sodium (VSVG) and potassium channels (KCNQ2/3), as well as 186 kDa neurofascin, a L1CAM cell adhesion molecule. Giant ankyrin-G also binds to and recruits cytoplasmic AIS molecules including beta-4-spectrin, and the microtubule-binding proteins, EB1/EB3 and Ndel1. Giant ankyrin-G is sufficient to rescue AIS formation in ankyrin-G deficient neurons. Ankyrin-G also includes a smaller 190 kDa isoform located at dendritic spines instead of the AIS, which is incapable of targeting to the AIS or rescuing the AIS in ankyrin-G-deficient neurons. Here, we described a protocol using cultured hippocampal neurons from ANK3-E22/23-flox mice, which, when transfected with Cre-BFP exhibit loss of all isoform of ankyrin-G and impair the formation of AIS. Combined a modified Banker glia/neuron co-culture system, we developed a method to transfect ankyrin-G null neurons with a 480 kDa ankyrin-G-GFP plasmid, which is sufficient to rescue the formation of AIS. We further employ a quantification method, developed by Salzer and colleagues to deal with variation in AIS distance from the neuronal cell bodies that occurs in hippocampal neuron cultures. This protocol allows quantitative studies of the de novo assembly and dynamic behavior of AIS.
The axon initial segment is located at the proximal axon in most vertebrate neurons. Functionally, AIS is where action potentials are initiated due to the high-density of voltage-gated sodium channels in this region. AIS of some excitatory neurons are also targeted by inhibitory interneurons through forming GABAergic synapses1,2,3. Therefore, AIS is a critical site to integrate cell signaling and modulate the excitability of neurons. AIS is normally 20-60 μm in length and located within 20 μm of the cell body. The length and position of AIS varies in neurons across brain regions, as well as in different developmental stages of the same neuron4,5. Accumulated evidence suggested that the composition and position of AIS are dynamic in responding to the change of neuronal activity4,5,6,7.
480 kDa ankyrin-G is the master organizer of AIS. 480 kDa ankyrin-G is a membrane associated adaptor protein that directly binds to voltage gated sodium channels as well as other major AIS proteins including beta4-spectrin, KCNQ2/3 channels that modulate sodium channel activity8,9, and 186 kDa neurofascin, a L1CAM that directs GABAergic synapses to the AIS2,10. 480 kDa ankyrin-G shares canonical ankyrin domains found in the short 190 kDa ankyrin-G isoform (ANK repeats, spectrin binding domain, regulatory domain), but are distinguished by a giant exon that is found only in vertebrates and is specifically expressed in neurons (Figure 1A)11,12. The 480 kDa ankyrin-G neuron specific domain (NSD) is required for AIS formation12. The 190 kDa ankyrin-G does not promote AIS assembly or target AIS in ankyrin-G-null neurons12. However, 190 kDa ankyrin-G is concentrated at the AIS containing 480 kDa ankyrin-G12. This ability of the 190 kDa ankyrin-G to target pre-assembled AIS of wildtype neurons has been a source of confusion in the literature and has slowed appreciation of the critical specialized functions of the 480 kDa ankyrin-G in AIS assembly. Therefore, it is critical to study AIS assembly in ankyrin-G-null neurons that lack a pre-assembled AIS.
Here, we present a method to study the assembly and structure of the AIS using cultured hippocampal neurons from ANK3-E22/23-flox mice that eliminates all isoforms of ankyrin-G13 (Figure 1B). By transfecting neurons with a Cre-BFP construct before AIS is assembled, we generated ankyrin-G-deficient neurons completely lacking an AIS (Figure 1B, Figure 2). The assembly of AIS is fully rescued following co-transfection of 480 kDa ankyrin-G-GFP plasmid with a Cre-BFP plasmid. This method provides a way to study the AIS assembly in a non-pre-assembled AIS environment. We also modified the glia-neuron co-culture system from Gary Banker without using antibiotics, previously designed for embryonic day 18 neurons, for application to postnatal mouse neurons and adapted a AIS quantitation method to average AIS measurements from multiple neurons to normalize the variation of AIS14,15.
NOTE: This culture method of hippocampal neurons from postnatal 0-day ANK3-E22/23f/f mice is adapted from Gary Banker’s glia/neuron co-culture system. Therefore, it is critical to perform all steps after dissection in a clean hood using sterilized tools. This protocol takes up to 1 month. The workflow is displayed in Figure 3. The protocol follows the animal guidelines of Duke University.
1. Preparing of coverslips and neuronal plating dishes
2. Preparing glia cell feeder dishes (2 weeks before culture day)
3. Culture hippocampal neurons
NOTE: All steps are performed at room temperature.
4. Disruption of AIS by Knockout of Ankyrin-G at earlier stage of neuron development
5. Quantification of axon initial segment
A complete set of experiment should include Cre-BFP only transfection as negative control, Cre-BFP plus 480 kDa ankyrin-G co-transfection as positive control and a non-transfected condition as technique control. In Cre-BFP only control, transfected neurons lack the accumulation of AIS markers, including ankyrin-G (ankG), beta4-spectrin (β4), neurofascin (Nf) and voltage gated sodium channels (VSVG) (Figure 4A)16. In contrast, Cre and 480 kDa ankyrin-G co-transfected neurons have fully assembled AIS revealed by the present of AIS markers (Figure 4B). It is important to confirm the quality of culture by comparing with the non-transfected dishes. Unhealthy neurons tend to show abnormal AIS structure, like discontinued or ectopic AIS (Figure 4C).
Then we showed an example of evaluating how an ankyrin-G human neurodevelopmental disorder mutation (ankG-K2864N) affects AIS assembly (Figure 5). 3 div ANK3-E22/23f/f neurons were transfected with Cre-BFP and wildtype 480 kDa ankyrin-G (ankG-WT) or 480 kDa ankyrin-G baring human mutation (ankG-K2864). Neurons were fixed at div7 and stained for ankyrin-G. Images were collected from 10-15 transfected neurons and 10-15 control neurons on the same coverslips and processed with maximum intensity projection. Then we draw a line at the AIS as shown and measure the mean intensity across the line. After averaging the AIS intensity, we plot the AIS intensity from the soma to the distal axon. AIS enriched protein normally showed a fast increase of signal from the proximal axon and a slow decrease of signal to the distal axon. AIS assembled by ankyrin-G with human mutant showed an increase and decrease of signal. But when aligned with the non-transfected AIS, the mutant curve is wider, and peak of the curve is lower suggesting a structure change of AIS. The wild type ankyrin-G assembled AIS closely aligned with the non-transfected one.
Figure 1: The genomic editing of ANK3-E22/23-flox. (A) Schematic representation of protein domains for 3 ankyrin-G isoforms. The location of exon 22 and 23 encoded regions in canonical domain is pointed by the dash line. (B) The position of LoxP sites in ANK3-E22/23-flox mice is indicated by triangle. In the present of Cre recombinase, exon 22 and 23 is deleted and causes loss the expression of all 3 isoforms of ankyrin-G. Please click here to view a larger version of this figure.
Figure 2: Loss of AIS in ANK3-E22/23-flox neurons in the present of Cre recombinase. A diagram shows the time frame of ankyrin-G expression and AIS assembly in wild type neurons versus in ANK3-E22/23f/f neurons with Cre transfection at 3 div. Please click here to view a larger version of this figure.
Figure 3: Workflow of protocol. Please click here to view a larger version of this figure.
Figure 4: A full rescue of AIS by 480kDa AnkG in ANK3-E22/23-flox neurons transfected with Cre. 3 div neurons of ANK3-E22/23f/f mice were transfected with Cre-BFP (A) or with a Cre-BFP and wild type 480 kDa ankyrin-G-GFP (B). Neurons were fixed at 7 div and stained for ankyrin-G (ankG), β4-spectrin (β4), neurofascin (Nf) and voltage gated sodium channels (VSVG). White arrow head points to the AIS of a transfected neuron. Scale bar is 20 μm. This figure was adapted from Yang et al16. (C) Two unhealthy neurons transfected with tdTM and 480 kDa ankyrin-G-GFP were shown. The formation of aggregates (circled in white and enlarged) is a sign of unhealthy neurons. Top: 480 kDa ankyrin-G shows up at the non-AIS region (pointed by white arrow heads. Bottom: Neuron formed 3 AIS and ectopic accumulation of ankyrin-G on soma. Scale bar is 20 μm. Please click here to view a larger version of this figure.
Figure 5: Quantification of AIS structural change. div3 neurons of ANK3-E22/23f/f mice were transfected with Cre-BFP and wild type 480 kDa ankyrin-G or 480 kDa ankyrin-G-K2864N. At 7 div, neurons were fixed and stained for ankyrin-G. Representative images show the ankyrin-G signal at the AIS. The green line and yellow line indicate where the line for AIS intensity measurement was drew. White dash line circled the cell body of the transfected neuron. Scale bar is 20 μm. Average AIS intensity for both conditions is plotted alone distance and aligned with non-transfected cells (n=10). Please click here to view a larger version of this figure.
The assembly of AIS is organized by 480 kDa ankyrin-G. However, ankyrin-G has shorter isoforms that can target to the AIS of wildtype neurons, which may lead to difficulty in interpretation of structure-function analyses of AIS assembly. Here we present a method using neurons from ANK3-E22/23-flox mice that allows study of de novo assembly of the AIS. By transfecting with Cre-BFP at 3 div, we eliminate the all endogenous isoforms of ankyrin-G. We could also co-transfect 480 kDa ankyrin-G to rescue the formation of AIS. This allows study of AIS formation in a clean system. By further adopting the Banker culture system which improves viability without complications of glial cell over-growth, we could reach a high transfection efficiency, which provide us enough neurons for quantitative measurement of AIS dimensions.
There are several critical steps in this protocol. The first critical step is considering the best time window to do the transfection, which needs to be early enough to prevent the assembly of AIS and late enough to reach the highest transfection efficiency. We tried 0 div electroporation transfection, which gave about 10% transfection efficiency with Cre-BFP only, but we were never able to transfect 480 kDa Ankyrin-G at 0 div. We suspect it is due to the large size of the plasmid (about 20 kb). Primary cultured hippocampal neurons have a narrow window for transfection, which is between 3-5 days. The accumulation of ankyrin-G at the AIS starts from 3 div. When we transfect Cre-BFP at 3 div, no AIS formation was seen in transfected neurons (Figure 4A). We could get 10-20 neurons transfected with 480 kDa ankyrin-G from one 18 mm coverslip. Also, for the co-transfection rescue experiment, all DNA must be generated under the same promoter and the ratio of Cre-BFP and 480 kDa ankyrin-G-GFP must be matched. In this experiment, we used chicken beta-actin promoter.
Another critical step is the modification to the Banker culture. The Banker culture was developed for culturing embryonic rat neurons. To better support the more sensitive mouse postnatal hippocampal neuron, we include a step of chopping hippocampi into smaller pieces to improve the trypsinization efficiency. Adding KOH treatment after the nitric acid treatment further reduced the toxicity from the glass coverslips, which help neurons attach and grow better.
A remaining challenge is how to control the expression level of ankyrin-G. A dosage screen helped to determine the optimal amount of plasmid used for transfection. Going forward, it is better to use a neuron-specific promoter to control the level of expression. The current data analysis did not measure the position of AIS. This function should be included in the future.
The authors have nothing to disclose.
We thank Dr. Gary Banker for suggestion on neuronal culture protocol. This work is supported by the Howard Hughes Medical Institute, a grant from NIH, and a George Barth Geller endowed professorship (V.B.).
10xHBSS | Thermo Fisher Scientific | 14065-056 | |
18mm coverglass (1.5D) | Fisher Scientific | 12-545-84-1D | |
190kDa ankyrin-G-GFP | Addgene | #31059 | |
2.5% Tripsin without phenol red | Thermo Fisher Scientific | 14065-056 | |
480kDa ankyrin-G-GFP | lab made | Provide upon request | |
ANK3-E22/23f/f mice | JAX | Stock No: 029797 | B6.129-Ank3tm2.1Bnt/J; |
B27 serum-free supplement | Thermo Fisher Scientific | A3582801 | |
Boric acid | Sigma-Aldrich | B6768 | |
Cell strainer with 70-mm mesh | BD Biosciences | 352350 | |
Ceramic coverslip-staining rack | Thomas Scientific | 8542E40 | |
Cre-BFP | Addgene | #128174 | |
D-Glucose | Sigma-Aldrich | G7021 | |
DMEM | Thermo Fisher Scientific | 11995073 | |
GlutaMAX-I supplement | Thermo Fisher Scientific | A1286001 | |
Lipofectamine 2000 | Thermo Fisher Scientific | 11668030 | |
MEM with Earle’s salts and L-glutamine | Thermo Fisher Scientific | 11095-080 | |
Neurobasal Medium | Thermo Fisher Scientific | 21103-049 | |
Nitric acid 70% | Sigma-Aldrich | 225711 | |
Opti-MEM I Reduced Serum Medium | Thermo Fisher Scientific | 31985062 | |
Paraformaldehyde | Sigma-Aldrich | P6148 | |
Penicillin-streptomycin | Thermo Fisher Scientific | 15140122 | |
Poly-L-lysine hydrochloride | Sigma-Aldrich | 26124-78-7 | |
Potassium hydroxide | Sigma-Aldrich | 1310-58-3 |