Deletion of Stk11 and Fos in BLA projection neurons alters intrinsic excitability and impairs formation of long-term aversive memory

Conditioned taste aversion (CTA) is a form of one-trial learning dependent on basolateral amygdala projection neurons (BLApn). Its underlying cellular and molecular mechanisms are poorly understood, however. We used RNAseq from BLApn to identify learning-related changes in Stk11, a kinase with well-studied roles in growth, metabolism and development, but not previously implicated in learning. Deletion of Stk11 restricted to BLApn completely blocks memory when occurring prior to training, but not following it, despite altering neither BLApn-dependent encoding of taste palatability in gustatory cortex, nor transcriptional activation of BLApn during training. Deletion of Stk11 in BLApn also increases their intrinsic excitability. Conversely, BLApn activated by CTA to express the immediate early gene Fos had reduced excitability. BLApn knockout of Fos also increased excitability and impaired learning. These data suggest that Stk11 and Fos expression play key roles in CTA long-term memory formation, perhaps by modulating the intrinsic excitability of BLApn. (149 words)

(also known as LKB1; Figure 2B) a kinase well studied in the context of cancer, cell 169 growth and development, but not previously known to be involved in learning and         which key exons are flanked by lox-p sites. In both cases, recombination leads to a 251 functionally null allele and analyses were carried out in homozygous ( f/f ) animals. Cre was 252 delivered by injecting AAV2/5-Camk2α::Cre-GFP into the BLA bilaterally. As expected, carrying the same genotypes but receiving AAV2/5-Camk2α::GFP served as controls.

256
Injections were performed 10 days prior to analysis to allow time for LOF to occur.

258
To assess the efficacy of this approach for CTA learning, we first examined the necessity  with Cre or control viruses 10 days before CTA training and were tested 48 hours later. Despite 300 significant reduction in saccharin consumption between testing and training sessions in control 301 mice, Stk11 KO animals showed almost no learning. but KO mice consumed 78% and this difference was highly significant (t(11)=-4.91; p=4E-4). (C) 308 Reduced saccharin consumption does not reflect overall reduction in drinking measured 8 hours  figure supplement 1).

318
Both Cre-GFP and GFP injected mice developed CTA ( Figure 4D)there was no 319 significant between-group difference in the intensity of memory, assessed from the ratio 320 of saccharin consumed during the test ( Figure 4E). Since the same deletion produces a 321 profound effect when occurring prior to training, this suggests that deletion of Stk11 from 322 BLApn does not affect the retention and retrieval of CTA memory, provided memory was 323 already formed prior to performing the knockout. This argues that Stk11 is required for 324 CTA memory formation.

327
Since Stk11 deletion must occur prior to CTA training to have an effect on learning, we 328 needed to rule out the possibility that its effect on memory came via disruption of either 329 the responsiveness of BLA neurons to training stimuli, or the output of BLA neurons, 330 which is known to be required for palatability coding within GC.

331
To assess the responsiveness of BLA neurons to training stimuli, we asked whether   (i.e. the rheobase was lower; Figure 6C). This reflected a higher resting input resistance 411 and a slightly lower voltage threshold. Some other electrophysiological properties also 412 differed (see Table 4) including sag ratios, action potential amplitudes, and the medium     BLApn promotes memory formation, rather than memory maintenance or retrieval, but it 576 remains difficult to entirely rule out a role for Stk11 prior to training, rather than during the 577 training process itself. We found that after Stk11 deletion, activation of Fos by training in 578 the BLA was not impaired, implying that at least the initial stages of transcriptional 579 activation leading to learning are intact. We also know that the ability of the BLA to convey 580 palatability information to the GC remained intact after Stk11 knockout. Prior studies have 581 shown that silencing of BLA neurons or of their axons within the GC impair palatability 582 coding during the late phase of GC gustatory responses. We found that these responses 583 were still present in GC following knockout, implying that this critical function of the BLA 584 for CTA learning remained intact.         Hyperpolarization activated sag was measured from responses to -100 pA current steps.

761
Action potential threshold, amplitude, afterhyperpolarization (AHP) and full width at half-762 height were averaged from the 5th-10th action potentials in trials with 10 to 20 Hz firing 763 rates. Action potential threshold is the membrane potential at which the slope first 764 exceeds 10 V/s. Amplitude was the difference between the peak value and the threshold 765 value of the action potentials. Sag ratio is defined as the ratio between maximum voltage 766 and steady-state voltage change with the negative current input. Medium AHP was 767 measured as membrane potential hyperpolarization after the action potentials mentioned 768 above compared to the action potential threshold. The slow AHP was measured from the 769 peak hyperpolarization following positive current steps generating 10-20 Hz firing.