Enhancement in motor learning through genetic manipulation of the Lynx1 gene

Abstract

The cholinergic system is a neuromodulatory neurotransmitter system involved in a variety of brain processes, including learning and memory, attention, and motor processes, among others. The influence of nicotinic acetylcholine receptors of the cholinergic system are moderated by lynx proteins, which are GPI-anchored membrane proteins forming tight associations with nicotinic receptors. Previous studies indicate lynx1 inhibits nicotinic receptor function and limits neuronal plasticity. We sought to investigate the mechanism of action of lynx1 on nicotinic receptor function, through the generation of lynx mouse models, expressing a soluble version of lynx and comparing results to the full length overexpression. Using rotarod as a test for motor learning, we found that expressing a secreted variant of lynx leads to motor learning enhancements whereas overexpression of full-length lynx had no effect. Further, adult lynx1KO mice demonstrated comparable motor learning enhancements as the soluble transgenic lines, whereas previously, aged lynx1KO mice showed performance augmentation only with nicotine treatment. From this we conclude the motor learning is more sensitive to loss of lynx function, and that the GPI anchor plays a role in the normal function of the lynx protein. In addition, our data suggests that the lynx gene plays a modulatory role in the brain during aging, and that a soluble version of lynx has potential as a tool for adjusting cholinergic-dependent plasticity and learning mechanisms in the brain.

Figure 1. Summary of three transgenic lynx1 constructs: L7 lynx1, L7 secreted lynx1, and lynx1 BAC.

(A) L7-lynx1 construct, upper, utilizing a pcp (L7) promoter sequence in front of a full length lynx1 cDNA contained the full length lynx1 coding sequence. Lower, L7-sec-lynx1, pcp (L7) promoter driving expression of a secreted variant of lynx1, lacking the final asparagine residue that makes up the mature form of lynx1, and lacking the GPI-anchor hydrophobic consensus sequence. An in-frame HA sequence replaces the GPI anchor sequence. (B) Representative structural model of lynx1 (right hand molecule) with associated GPI-linked tether (red circles) to the plasma membrane (grey). Model is based on the NMR structure of lynx1 . Left hand molecule, the secreted variant of lynx1 can translocate freely across the plasma membrane and diffuse into the extracellular and/or synaptic space. (C) BAC modification strategy for the generation of a lynx1 modified BAC transgenic mouse line.

Figure 2. Rotarod performance for L7-sec-lynx1 transgenic mice.

(A) Line 1. No differences in performance were observed initially, but the transgenic mice (L7-sec-lynx1 Tg) outperformed their wild-type littermates beginning at the 5^th trial. These data indicate that there are no motor performance differences in the transgenic mice vs. wildtype mice, but that there is a significant improvement in motor learning behavior. Y axis is in seconds; X axis is trial number. P<0.05. (B) Line 1. Motor performance over five consecutive training days. Average of all trials per day indicates a significant performance enhancement in L7-sec-lynx1 Tg as compared to their wild-type littermate controls. Y axis is in seconds, X axis is in days. (C) Line 2. Motor performance using a slow acceleration rotarod paradigm (0.1 RPM/sec). Significant enhancements in rotarod performance in L7-sec-lynx1 Tg mice were observed in latter trials. Y axis is in seconds; X axis is in trials. P<0.05. (D) Table of data collected for two independent founder lines of the L7-sec-lynx1 transgene tested on an accelerating rotarod paradigm, with two separate acceleration paradigms of the rotarod test (modes) used, slower (0.1 RPM/sec.) and faster (1 RPM/sec.) accelerating paradigm. In both paradigms, both L7-lynx1 transgenic mouse lines demonstrated a significant enhancement in motor learning but no differences in baseline motor performance. (E) Summary of the effects of motor learning on the two lines of the same L7-sec-lynx1 transgenic mice. The maximal daily performance from all training days is plotted as a percentage relative to wild-type. “f” and “s” suffixes refer to the faster or slower acceleration, respectively.

Figure 3. Comparison of motor learning effects across several genetically modified lynx lines.

Overexpression lines, L7-lynx1 and lynx1BAC mice, which over-expressed full length lynx1 in Purkinje cells, and lynx1-expressing cells, respectively, did not display significant differences in motor learning. (A) L7-lynx1 transgenic mice, overexpressing full length lynx1 in Purkinje cells, showed no differences in motor performance or learning as compared to their wild-type counterparts. Y axis is in seconds, X is in trials. (B) L7-lynx1 transgenic mice exhibited no learning improvements over that of wild-type littermates. Y axis is in seconds, X axis is in days. (C) lynx1 BAC transgenic mice showed no difference in motor performance as compared to wild-type littermates in initial training trials on the rotarod. Y axis is in seconds, X is in trials. No significant differences were observed. (D) On subsequent training days, lynx1BAC mice did not show a difference in motor learning. Y axis is in seconds. No significant differences were observed. (E) Comparisons of motor learning across lines. Data is represented for each line as a percentage over the performance of their wild-type counterparts. The improvement in motor learning of L7-sec-lynx1 transgenic mice contrast with expression of full-length lynx1 using lynx1BAC transgenesis. Rotarod motor learning was also sensitive to removal of lynx (lynx1 KO), and demonstrated a significant increase over wild-type controls. Aged mice, (>1 yr), showed no motor enhancements unless treated with nicotine . Y axis is percentage over wild-type controls. Data depicted represent the seventh training trial for each group.

Figure 4. Two schemes for possible modes of action of secreted lynx on nicotinic receptors within synapses.

(A) Schematic diagram of a synapse in WT mice. Schematic model of nicotinic receptors and lynx1 interaction at a Purkinje cell synapse. Models are based on the crystal structure of AChBP and the NMR structure of lynx1 . Lynx1 is depicted as binding at the subunit interface of the pentameric channel, based on α-bungarotoxin binding. Lynx1 expression is expressed in the post-synaptic cell (Purkinje cell), and is not expressed in the presynaptic neuron (stellate/basket neuron). Tethered to the membrane by it GPI-anchor, lynx is depicted as having access to a nicotinic receptor binding site at the post-synaptic face only. (B) Schematic model 1 of sec-lynx1 function – cell autonomous, dominant negative model. Schematic representation of Purkinje cell synapses in L7-sec-lynx1 Tg mice. Binding of secreted lynx1 to the same subunit interface of the nicotinic receptor could compete off the binding of native full-length GPI-anchored lynx1, and thereby exert a dominant negative effect. This model implies that the secreted version of lynx1 has either no effect or a differential function as compared to the membrane bound version of lynx1, but maintains nicotinic receptor binding capability. (C) Schematic model 2 of sec-lynx1 function – circuit based, ectopic expression model. In this model, the soluble lynx1 secreted from the post-synaptic Purkinje cells diffuses extracellularly, accessing nicotinic receptors located on terminals of pre-synaptic neurons (stellate/basket cell). In this model, the ectopic expression of lynx1 in pre-synaptic sites can lead to suppression of activity or neurotransmitter release, leading to dis-inhibition onto Purkinje cells. This dis-inhibition can lead to alterations in excitatory/inhibitory balance and motor learning.