Amyloid β-Peptide Causes the Permanent Activation of CaMKIIα through Its Oxidation

Alzheimer’s disease (AD) is characterised by the presence of extracellular amyloid plaques in the brain. They are composed of aggregated amyloid beta-peptide (Aβ) misfolded into beta-sheets which are the cause of the AD memory impairment and dementia. Memory depends on the hippocampal formation and maintenance of synapses by long-term potentiation (LTP), whose main steps are the activation of NMDA receptors, the phosphorylation of CaMKIIα and the nuclear translocation of the transcription factor CREB. It is known that Aβ oligomers (oAβ) induce synaptic loss and impair the formation of new synapses. Here, we have studied the effects of oAβ on CaMKIIα. We found that oAβ produce reactive oxygen species (ROS), that induce CaMKIIα oxidation in human neuroblastoma cells as we assayed by western blot and immunofluorescence. Moreover, this oxidized isoform is significantly present in brain samples from AD patients. We found that the oxidized CaMKIIα is active independently of the binding to calcium/calmodulin, and that CaMKIIα phosphorylation is mutually exclusive with CaMKIIα oxidation as revealed by immunoprecipitation and western blot. An in silico modelling of the enzyme was also performed to demonstrate that oxidation induces an activated state of CaMKIIα. In brains from AD transgenic models of mice and in primary cultures of murine hippocampal neurons, we demonstrated that the oxidation of CaMKIIα induces the phosphorylation of CREB and its translocation to the nucleus to promote the transcription of ARC and BDNF. Our data suggests that CaMKIIα oxidation would be a pro-survival mechanism that is triggered when a noxious stimulus challenges neurons as do oAβ.


Introduction
Alzheimer's disease (AD) is a neurodegenerative disease with the highest prevalence worldwide [1]. AD is a type of dementia whose main histopathological hallmarks are the presence of amyloid β-peptide (Aβ) forming extracellular senile plaques and intracellular fibrillary tangles caused by tau aggregation [2]. High concentrations of Aβ aggregate forming neurotoxic oligomers, initiating the damage in the hippocampus and causing memory loss [3,4].
The hippocampus is the centre in the brain for memory and learning based on glutamatergic neurotransmission. Glutamate released by the presynaptic terminal binds to both NMDA and AMPA receptors [5] at the postsynaptic terminal, allowing the entrance of sodium and calcium into the postsynaptic neuron. An increase in synaptic transmission will lead to a sustained calcium entry, which is necessary to induce long-term potentiation

Oxidation Induces Autonomous CaMKIIα Activity
It was previously reported that in cardiomyocytes ROS causes the pathophysiological activation of CaMKII by its oxidation [12,13], an activation state that occurs after Ca 2+ /CaM binding and lasts after the release of the Ca 2+ /CaM ( Figure 1A). Here we have studied if this oxidation occurs in the α isoform in vitro ( Figure 1B-D). Firstly, we demonstrated that CaMKIIα activity is dependent on the previous binding of Ca 2+ /CaM, since there was no activity when Ca 2+ /CaM was not preincubated with the enzyme ( Figure 1B). Moreover, the oxidative stress inducer H 2 O 2 did not affect the activity of CaMKIIα when it was bound to Ca 2+ /CaM ( Figure 1B). Then, the activated enzyme was incubated with the calcium chelator BAPTA-AM previous to any contact with ATP, demonstrating that once the enzyme was oxidized it remained active even in the absence of Ca 2+ [10,14] (p < 0.05) ( Figure 1C). Finally, the incubation of the enzyme with the classical CaMKIIα inhibitor Kn-93 was able to reduce native CaMKIIα activity (p < 0.001), but not the activity of oxidized CaMKIIα ( Figure 1D).

Aβ Triggers CaMKIIα Oxidation
Aβ induces neuronal death mostly by oxidative stress [11,15]. Therefore, we aimed the study on the effect of oAβ 1-42 on neuroblastoma cells ( Figure 2). Cells were treated with oAβ 1-42 or H 2 O 2 , and the production of oxidative stress was analysed by the oxidation of H 2 DCFDA ( Figure 2A). As expected, 10 µM oAβ 1-42 and 200 µM H 2 O 2 produced statistically significant amounts of free radicals (p < 0.05 and p < 0.0001, respectively). Furthermore, we studied the toxic effect of oAβ 1-42 on neuroblastoma cells by MTT reduction assay ( Figure 2B). We found that 10 µM oAβ 1-42 was neurotoxic after 24 h of treatment (p < 0.0001) and this toxicity was partially reverted by the CaMKIIα inhibitor Kn-93 (p < 0.05), suggesting that a prolonged activation of CaMKIIα by oAβ 1-42 could have deleterious effects.
In order to study the effect of oAβ 1-42 on CaMKIIα, we pre-treated SH-SY5Y cells either with or without 10 µM of the calcium chelator BAPTA-AM for 30 min, then we added either 10 µM oAβ 1-42 or 200 µM H 2 O 2 , and cells were incubated for 30 min. Ox-CaMKIIα and the total CaMKIIα were studied by western blot (WB) (Figure 3). We found that both Representation of CaMKIIα activation mechanism. After Ca 2+ /CaM binding the enzyme turns active by classical phosphorylation but also by oxidation. Later the Ca 2+ /CaM is released, and the enzyme continues to be active by oxidation in a state termed autonomous. (B) The activity of the enzyme was assayed in vitro by a commercial kit showing that 10 μM H2O2 is unable to activate CaMKIIα without Ca +2 /CaM binding, neither is it able to affect the native activity of CaMKIIα. Data are the mean ± SEM of 5 independent experiments. non significant (ns) vs. the respective controls by twoway ANOVA plus Bonferroni as post-hoc test. (C) After initial activation by Ca +2 /CaM, 10 μM H2O2 can induce an active state, which is independent of Ca +2 /CaM as it was demonstrated by using 2 mM of the calcium chelator BAPTA-AM. Data are the mean ± SEM of 4 independent experiments. p < 0.05 vs. control by Student's t-test. (D) In vitro analysis of the CaMKIIα activity showed that the activity of ox-CaMKIIα cannot be inhibited by 1 μM of the specific inhibitor Kn-93. Data are the mean ± SEM of 10 independent experiments. ns, * p < 0.05, *** p < 0.001 vs. the respective controls by ANOVA plus Bonferroni as post-hoc test.

Aβ Triggers CaMKIIα Oxidation
Aβ induces neuronal death mostly by oxidative stress [11,15]. Therefore, we aimed the study on the effect of oAβ1-42 on neuroblastoma cells ( Figure 2). Cells were treated with oAβ1-42 or H2O2, and the production of oxidative stress was analysed by the oxidation of H2DCFDA ( Figure 2A). As expected, 10 μM oAβ1-42 and 200 μM H2O2 produced statistically significant amounts of free radicals (p < 0.05 and p < 0.0001, respectively). Furthermore, we studied the toxic effect of oAβ1-42 on neuroblastoma cells by MTT reduction assay ( Figure 2B). We found that 10 μM oAβ1-42 was neurotoxic after 24 h of treatment (p < 0.0001) and this toxicity was partially reverted by the CaMKIIα inhibitor Kn-93 (p < 0.05), suggesting that a prolonged activation of CaMKIIα by oAβ1-42 could have Representation of CaMKIIα activation mechanism. After Ca 2+ /CaM binding the enzyme turns active by classical phosphorylation but also by oxidation. Later the Ca 2+ /CaM is released, and the enzyme continues to be active by oxidation in a state termed autonomous. (B) The activity of the enzyme was assayed in vitro by a commercial kit showing that 10 µM H 2 O 2 is unable to activate CaMKIIα without Ca +2 /CaM binding, neither is it able to affect the native activity of CaMKIIα. Data are the mean ± SEM of 5 independent experiments. non significant (ns) vs. the respective controls by two-way ANOVA plus Bonferroni as post-hoc test. (C) After initial activation by Ca +2 /CaM, 10 µM H 2 O 2 can induce an active state, which is independent of Ca +2 /CaM as it was demonstrated by using 2 mM of the calcium chelator BAPTA-AM. Data are the mean ± SEM of 4 independent experiments. p < 0.05 vs. control by Student's t-test. (D) In vitro analysis of the CaMKIIα activity showed that the activity of ox-CaMKIIα cannot be inhibited by 1 µM of the specific inhibitor Kn-93. Data are the mean ± SEM of 10 independent experiments. ns, * p < 0.05, *** p < 0.001 vs. the respective controls by ANOVA plus Bonferroni as post-hoc test. Kn-93 is able to induce protection regarding oAβ-mediated oxidative stress. (A) SH-SY5Y cells were treated with 10 μM oAβ1-42 or 200 μM H2O2 for 2 h. ROS production was measured quantifying the oxidation of H2DCFDA by the fluorescence emission at 522 nm. Data are the mean ± SEM of 8 independent experiments. * p < 0.05, **** p < 0.0001 vs. controls by ANOVA plus Bonferroni as post-hoc test. (B) SH-SY5Y cells were treated with 10 μM oAβ1-42 for 24 h with or without 1 μM Kn-93. Data are the mean ± SEM of 4-8 independent experiments. non significant (ns), ### p < 0.001, #### p < 0.0001 vs. the respective controls by ANOVA plus Bonferroni as post-hoc test.
In order to study the effect of oAβ1-42 on CaMKIIα, we pre-treated SH-SY5Y cells either with or without 10 μM of the calcium chelator BAPTA-AM for 30 min, then we added either 10 μM oAβ1-42 or 200 μM H2O2, and cells were incubated for 30 min. Ox-CaMKIIα and the total CaMKIIα were studied by western blot (WB) (Figure 3). We found that both oAβ1-42 and H2O2 induce CaMKIIα oxidation (p < 0.001) ( Figure 3A). The same results were obtained regarding the effect of oAβ1-42 and H2O2 on CaMKIIα oxidation In order to demonstrate that within all CaMKII isoforms, the α isoform is able to become oxidized, we performed a PLA assay. This assay shows if two antigens recognized by two different antibodies (in this case anti-ox-CaMKII and anti-CaMKIIα) coexist within 50 nm, sharing a proximal location [16] ( Figure 3C). Our results demonstrate that neuroblastoma cells treated with 10 µM oAβ 1-42 for 30 min presented significant CaMKIIα oxidation (p < 0.001) ( Figure 3C).
Once it was demonstrated that CaMKIIα is oxidized by oAβ 1-42, we addressed the study of the relationship between both post-translational modifications, phosphorylation and oxidation. Cells were lysed, and phosphorylated CaMKIIα was immunoprecipitated in a proportion of 1:10 regarding the input (Figure 4). Multiple WBs were run to study the presence of total, phosphorylated and oxidized CaMKIIα, demonstrating that both modifications are mutually exclusive.
The pathophysiological relevance of these findings was assayed by the study of the CaMKIIα in AD brain samples ( Figure 5). We performed immunofluorescence analysis on cortical samples from AD patients compared to samples from non-demented donors.
As expected, we found a significant increase in CaMKIIα oxidation in AD brain samples compared to non-demented donors (p < 0.05). when cells were studied by immunofluorescence ( Figure 3B). Consistently, this oxidation was impaired when cells were pre-incubated with BAPTA-AM, which prevented the required activation by Ca 2+ /CaM, making it impossible to reach the autonomous state of activation ( Figure 3A,B). In order to demonstrate that within all CaMKII isoforms, the α isoform is able to become oxidized, we performed a PLA assay. This assay shows if two antigens recognized Cell extracts were obtained and analysed by WB. A representative WB is shown. Oxidized and total CaMKIIα bands were normalized regarding GAPDH and the quantifications are shown in the graph. Data are the mean ± SEM of 3 independent experiment. ** p < 0.005, *** p < 0.001 vs. the respective controls by ANOVA plus Bonferroni as post-hoc test. The uncropped gels are shown in Figure S3. (B) SH-SY5Y cells were treated as explained in (A). Immunofluorescence experiments were performed with Abs against ox-CamKII and CaMKIIα. Results were analysed by confocal microscopy and representative images are shown. Fluorescence intensity was measured and expressed as the ratio between ox-CaMKIIa and the total enzyme in the graph. Data are the mean ± SEM of 5 independent experiments. ** p < 0.005 vs. the respective controls by ANOVA plus Bonferroni as post-hoc test. (C) The specific oxidation of the CaMKIIα isoform was studied by PLA assay. SH-SY5Y cells were treated with 10 µM oAβ 1-42 for 30 min. A representative set of images is shown. The number of dots was quantified and normalized regarding the number of nuclei and shown in the right graph. Data are the mean ± SEM of 4 independent experiments. *** p < 0.001 vs. controls by Student's t-test. oxidation (p < 0.001) ( Figure 3C).
Once it was demonstrated that CaMKIIα is oxidized by oAβ1-42, we addressed the study of the relationship between both post-translational modifications, phosphorylation and oxidation. Cells were lysed, and phosphorylated CaMKIIα was immunoprecipitated in a proportion of 1:10 regarding the input ( Figure 4). Multiple WBs were run to study the presence of total, phosphorylated and oxidized CaMKIIα, demonstrating that both modifications are mutually exclusive.

Figure 4.
Phosphorylation and oxidation of CaMKIIα are mutually exclusive. SH-SY5Y cells were lysed, and the extracts were immunoprecipitated with an anti p-Thr286-CaMKIIα Ab. Then, immunoprecipitates were analysed by WB with Abs against pCaMKIIα, total CaMKIIα and ox-CaMKII. An input of the extracted protein was used as control, while GAPDH was used as control of the immunoprecipitation efficiency. WB did not show ox-CaMKII signal, meaning that oxidation was not occurring in the pCaMKIIα immunoprecipitates. A representative set of experiments is shown from 3 independent experiments. The respective negative controls and the WB obtained in the different experiments are shown in Figure S4.
The pathophysiological relevance of these findings was assayed by the study of the CaMKIIα in AD brain samples ( Figure 5). We performed immunofluorescence analysis on cortical samples from AD patients compared to samples from non-demented donors. As expected, we found a significant increase in CaMKIIα oxidation in AD brain samples compared to non-demented donors (p < 0.05). . Phosphorylation and oxidation of CaMKIIα are mutually exclusive. SH-SY5Y cells were lysed, and the extracts were immunoprecipitated with an anti p-Thr286-CaMKIIα Ab. Then, immunoprecipitates were analysed by WB with Abs against pCaMKIIα, total CaMKIIα and ox-CaMKII. An input of the extracted protein was used as control, while GAPDH was used as control of the immunoprecipitation efficiency. WB did not show ox-CaMKII signal, meaning that oxidation was not occurring in the pCaMKIIα immunoprecipitates. A representative set of experiments is shown from 3 independent experiments. The respective negative controls and the WB obtained in the different experiments are shown in Figure S4.
ci. 2022, 23, x FOR PEER REVIEW 8 of 18 Figure 5. Study of the oxidation of CaMKIIα in cortical brain samples from non-demented donors and AD patients. Immunofluorescence experiments were performed with Abs against ox-CamKII and CaMKIIα. Nuclei were stained with Dapi. Scale bar represents 100 μM. Results were analysed by confocal microscopy and representative images are shown. Fluorescence intensity was measured and expressed as the ratio between ox-CaMKII and CaMKIIα in the right graph. Data are the mean ± SEM of 3-4 independent experiments. * p < 0.05 vs. the non-demented donors by Student t-test.

In Silico Modelling of ox-CaMKIIα
We modelled the structure of CaMKIIα in silico and we also modelled the effect of putative mutants that mimicked the effect of the oxidation at C280 and/or M281 ( Figure  6). An amplified detail of the loop shows the potential interaction between phosphorylated Thr-286 and Arg-283 in order to stabilize the helix capping ( Figure 6A). We studied the effect of mutant isoforms by changing the phosphorylated and oxidized amino acids by glutamic ( Figure 6B), which mimics the effect of such post-translational modifications resulting in the mutants T286E and C280E/M281E. The reliability of the Figure 5. Study of the oxidation of CaMKIIα in cortical brain samples from non-demented donors and AD patients. Immunofluorescence experiments were performed with Abs against ox-CamKII and CaMKIIα. Nuclei were stained with Dapi. Scale bar represents 100 µM. Results were analysed by confocal microscopy and representative images are shown. Fluorescence intensity was measured and expressed as the ratio between ox-CaMKII and CaMKIIα in the right graph. Data are the mean ± SEM of 3-4 independent experiments. * p < 0.05 vs. the non-demented donors by Student t-test.

In Silico Modelling of ox-CaMKIIα
We modelled the structure of CaMKIIα in silico and we also modelled the effect of putative mutants that mimicked the effect of the oxidation at C280 and/or M281 ( Figure 6). An amplified detail of the loop shows the potential interaction between phosphorylated Thr-286 and Arg-283 in order to stabilize the helix capping ( Figure 6A). We studied the effect of mutant isoforms by changing the phosphorylated and oxidized amino acids by glutamic ( Figure 6B), which mimics the effect of such post-translational modifications resulting in the mutants T286E and C280E/M281E. The reliability of the prediction is scored from 0 to 9, 9 being the most reliable prediction (indicated as Conf). The prediction shows that phosphorylated T286 can help to sustain the helix (stabilization of the capping) despite extending the loop, while the oxidation of C280 and Met281 will extend the helix. We hypothesise that after oxidation of C280 and M281, the conformation of CaM-binding helix in the N-terminal domain of calmodulin-kinase will be less flexible, being more difficult to recover its original structure ( Figure 6C). On the other hand, the phosphorylation of Thr286 will help to preserve the original conformation of the helix and even increase the flexibility of the hinge loop, helping to recover the original structure after the clamp of calmodulin releases the binding helix ( Figure 6C).

Pathophysiological Effects of ox-CaMKIIα
The CaMKII-CREB pathway is known to be protective for neurons since CREB phosphorylation and its nuclear translocation will start the transcription of pro-survival genes [18,19]. oAβ1-42 is a well-known inducer of neuronal toxicity [20][21][22] and one of the main mechanisms of neurotoxicity is the production of oxidative stress [23][24][25], which will promote the oxidation of CaMKIIα. Consistently, we have found that transgenic mice overexpressing APPswe/PSEN1dE9 ( Figure 7A) have a significant higher expression of p-CREB than the cortical samples from wild type mice (p < 0.0001). The presence of Aβ deposits in the cortex of the mice is showed in Figure S1. Therefore, we have studied the role of oAβ1-42 on CREB phosphorylation ( Figure 7B). Our results demonstrated that oAβ1-42 induced a prolonged activation of CREB by its phosphorylation (p < 0.01). This effect

Pathophysiological Effects of ox-CaMKIIα
The CaMKII-CREB pathway is known to be protective for neurons since CREB phosphorylation and its nuclear translocation will start the transcription of pro-survival genes [18,19]. oAβ 1-42 is a well-known inducer of neuronal toxicity [20][21][22] and one of the main mechanisms of neurotoxicity is the production of oxidative stress [23][24][25], which will promote the oxidation of CaMKIIα. Consistently, we have found that transgenic mice overexpressing APPswe/PSEN1dE9 ( Figure 7A) have a significant higher expression of p-CREB than the cortical samples from wild type mice (p < 0.0001). The presence of Aβ deposits in the cortex of the mice is showed in Figure S1. Therefore, we have studied the role of oAβ 1-42 on CREB phosphorylation ( Figure 7B). Our results demonstrated that oAβ 1-42 induced a prolonged activation of CREB by its phosphorylation (p < 0.01). This effect was independent of the phosphorylation state of CaMKIIα, since the inhibitor Kn-93 did not produce effects on the levels of p-CREB (p < 0.001). The downstream effect of CREB phosphorylation is the activation of the transcription of genes related with memory and learning and with neuronal survival [18,19]. For this reason, we have analysed by RT-PCR the transcription of ARC ( Figure 8A) and BDNF ( Figure 8B). In both cases (Figure 8), we found that the presence of Aß and the consequent oxidation of CaMKIIα increase the transcription of BDNF and ARC along 24 h independent of the presence of the CaMKIIα inhibitor Kn-93. (B) Primary cultures of mouse hippocampal neurons were treated with 1 µM Kn-93 and 10 µM oAβ 1-42 for 30 min. CREB phosphorylation was analysed by immunofluorescence with an Ab against p-S133-CREB. A representative set of images is shown. Fluorescence of p-CREB was quantified and the data are shown in the right graph. Data are the mean ± SEM of 6 different fields. ** p < 0.01, **** p < 0.0001 vs. controls by two-way ANOVA plus Bonferroni post-hoc test.
The downstream effect of CREB phosphorylation is the activation of the transcription of genes related with memory and learning and with neuronal survival [18,19]. For this reason, we have analysed by RT-PCR the transcription of ARC ( Figure 8A) and BDNF ( Figure 8B). In both cases (Figure 8

Discussion
Memory loss is a very early symptom of AD, which progresses to dementia. In particular, the hippocampus which is the centre of memory and learning, is dramatically affected by Aβ in AD. Therefore, we have studied the role of oAβ on the intracellular signalling that impairs memory formation. We focused our work in the post-translational modifications of CaMKIIα, the key enzyme that transmits the signal from the postsynaptic term to the soma allowing the translocation of CREB, a mechanism not just involved in memory formation but also in neuroprotection [26].
CaMKII oxidation has been studied previously in the heart, where it was found that it is pathological [10]. Consistently, we have found that CaMKIIα can be oxidized in vitro and this oxidized state of the enzyme results in permanent activity even in the absence of Ca 2+ . This state is termed "autonomous" since its activity becomes independent of Ca 2+ once it is oxidized. Therefore, our goal was to study of the effect of oAβ1-42 on CaMKIIα, since oxidative stress is playing a key pathophysiological role in AD [20]. oAβ1-42 is able to produce oxidative stress by itself [11,27], and here we have demonstrated that oAβ1-42 is producing free radicals that affect the cell viability of human neuroblastoma cells. The free radicals produced by oAβ1-42 oxidize CaMKIIα. This post-translational modification induces a CaMKIIα autonomous active state, like H2O2 does. The presence of ox-CaMKIIα in brain samples from AD patients show the pathophysiological relevance of the findings.
We have modelled the oxidation of CaMKIIα in silico to demonstrate that it produces the activation of the enzyme. This post-translational modification shows a different structural stability than CaMKIIα phosphorylation. As demonstrated by the structural studies that we performed, once Ca 2+ /CaM binds to CaMKIIα, the hinge structure that hides the catalytic domain opens. Both CM280-281 and T286 are located in the sequence that constitutes this hinge. Through CaMKIIα oxidation, this sequence becomes an αhelix, providing a structural stability that enables CaMKIIα to work in an autonomous way. Even though both CM280-281 oxidation and T286 phosphorylation stabilize the structure, they are mutually exclusive because their post-translational modifications share a common location in the α-helix. Furthermore, α-helix stability is lower in the oxidized state, which supports our enzymatic activity results. Finally, since oxidation remains for a longer time than phosphorylation, the final model depicts an ox-CaMKIIα that is mildly but constantly active compared to the phosphorylated CaMKIIα.
The capacity of sustaining an autonomous active state through phosphorylation enables CaMKIIα to change its conformational state and facilitate the phosphorylation of

Discussion
Memory loss is a very early symptom of AD, which progresses to dementia. In particular, the hippocampus which is the centre of memory and learning, is dramatically affected by Aβ in AD. Therefore, we have studied the role of oAβ on the intracellular signalling that impairs memory formation. We focused our work in the post-translational modifications of CaMKIIα, the key enzyme that transmits the signal from the postsynaptic term to the soma allowing the translocation of CREB, a mechanism not just involved in memory formation but also in neuroprotection [26].
CaMKII oxidation has been studied previously in the heart, where it was found that it is pathological [10]. Consistently, we have found that CaMKIIα can be oxidized in vitro and this oxidized state of the enzyme results in permanent activity even in the absence of Ca 2+ . This state is termed "autonomous" since its activity becomes independent of Ca 2+ once it is oxidized. Therefore, our goal was to study of the effect of oAβ 1-42 on CaMKIIα, since oxidative stress is playing a key pathophysiological role in AD [20]. oAβ 1-42 is able to produce oxidative stress by itself [11,27], and here we have demonstrated that oAβ 1-42 is producing free radicals that affect the cell viability of human neuroblastoma cells. The free radicals produced by oAβ 1-42 oxidize CaMKIIα. This post-translational modification induces a CaMKIIα autonomous active state, like H 2 O 2 does. The presence of ox-CaMKIIα in brain samples from AD patients show the pathophysiological relevance of the findings.
We have modelled the oxidation of CaMKIIα in silico to demonstrate that it produces the activation of the enzyme. This post-translational modification shows a different structural stability than CaMKIIα phosphorylation. As demonstrated by the structural studies that we performed, once Ca 2+ /CaM binds to CaMKIIα, the hinge structure that hides the catalytic domain opens. Both CM280-281 and T286 are located in the sequence that constitutes this hinge. Through CaMKIIα oxidation, this sequence becomes an α-helix, providing a structural stability that enables CaMKIIα to work in an autonomous way. Even though both CM280-281 oxidation and T286 phosphorylation stabilize the structure, they are mutually exclusive because their post-translational modifications share a common location in the α-helix. Furthermore, α-helix stability is lower in the oxidized state, which supports our enzymatic activity results. Finally, since oxidation remains for a longer time than phosphorylation, the final model depicts an ox-CaMKIIα that is mildly but constantly active compared to the phosphorylated CaMKIIα.
The capacity of sustaining an autonomous active state through phosphorylation enables CaMKIIα to change its conformational state and facilitate the phosphorylation of other monomers that constitute the physiological homododecameric CaMKIIα. This physiological effect is able to codify the intensity of neuronal activity and induce different kinds of signalling such as neuroprotection or the induction of LTP. Therefore, it is necessary to know if the effect of a pathologically driven oxidation of CaMKIIα is beneficial or detrimental for neurons. Since we found that CREB phosphorylation is increased in AD model mice and in hippocampal neurons challenged with oAβ 1-42 , we hypothesise that CaMKIIα oxidation would be a protective mechanism triggered by neurons after being challenged by oAβ 1-42 . This mechanism is a steady low active state, while phosphorylation is able to transduce potent calcium signalling [9,10,28,29]. This potent calcium signalling leads to an exponential threshold signalling necessary for the induction of LTP through phosphorylation of CREB at S133, S142 and S143 [30,31]. Thus, our results suggest a possible initial neuroprotective effect regarding the oxidative stress produced by oAβ 1-42 through CaMKIIα oxidation to maintain CREB activity, which induces the transcription of genes involved in synaptic plasticity and survival such as ARC and BDNF, as we have found.
In conclusion, in this work we report the existence of CaMKIIα oxidation in AD brain samples and in neuroblastoma cells treated with oAβ 1-42 as a model of AD. This oxidation generates a Ca/CaM autonomous state that may be beneficial regarding the beginning of Aβ neurotoxicity. However, due to the mutual exclusion of oxidation and physiological phosphorylation, this will lead to CaMKIIα becoming inactive. This inactivity can affect LTP and ultimately lead to neurotoxicity and cell death.

Mouse Cortical Brain Samples
Mouse cortical brain samples were obtained from B6C3-Transgenic (APPswePSEN1∆E9) 85Dbo/J mice purchased from Jackson Laboratory (Bar Harbor, ME, USA) and wild-type mice. The procedure was carried out according to the rules of the Helsinki Declaration and to the Comité Ético Científico para el Cuidado de Animales y Ambiente (CEC-CAA) from Pontifical Catholic University of Chile. Samples were obtained from 19-month-old animals.

Aβ Oligomerization
Aliquots of monomeric Aβ were maintained at 4 • C for 24 h without agitation to allow its aggregation. After 24 h, they were frozen at −20 • C until use.

CaMKIIα Kinase Assay
CaMKIIα activity was measured by a commercial kinase activity kit (Promega, Madison, WI, USA). Briefly, 1 ng of CaMKIIα activity was assessed by the kinase ATP consumption kit (ADP-Glo Promega) and inferred from a consumption curve of 25 µM ATP/ADP at different ratios. The kinase was previously treated to induce its oxidation with 10 µM H 2 O 2 , free Ca 2+ was chelated in some conditions using 2 mM BAPTA-AM (Millipore, Darmstadt, Germany) and the CaMKIIα inhibitor Kn-93 (MedChemExpress, Monmouth Junction, NJ, USA) was used at 1 µM. Then, the kit's protocol was followed.

Immunoprecipitation and Study of CaMKIIα Phosphorylation and Oxidation
Twenty-four hours before the protein extraction from SH-SY5Y cells, we prepared sepharose beads for immunoprecipitation. Briefly, 100 µL sepharose G beads (Fisher Scientific, Hampton, NH, USA) were washed thrice with ice-cold TBS and centrifuged in between (3000× g; 3 min). Then, the beads were incubated o.n. with 1:50 rabbit anti-ox-CaMKII Ab (GeneTex) or 1:50 anti-phosphorylated-Thr286 CaMKIIα Ab (Thermo Fisher, Waltham, MA, USA) in ice-cold TBS with protease inhibitor cocktail (Roche Basilea, Switzerland) in rotation at 4 • C. Then, the beads were centrifuged at 3000× g for 3 min and blocked with 1% BSA in TBS for 1 h in rotation at 4 • C. Beads were washed again thrice with ice-cold TBS and kept at 4 • C while the procedure was carried out.
SH-SY5Y cells seeded in a 75 cm 2 flask at 80% confluence were washed twice with PBS, trypsinized for 5 min and then centrifuged at 1200 rpm for 5 min. Pellets were lysed with 700 µL of non-denaturing lysis buffer (150 mM Sodium Chloride, 1.0% TritonX-100, 50 mM Tris, pH 8.0) supplemented with phosphatase and protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 mM N-ethylmaleimide and 1 mM sodium orthovanadate), and centrifuged at 15,000× g for 10 min at 4 • C. 250 µL of extracted protein were added to the bead-Ab complexes and incubated in rotation for 4 h at 4 • C. Next, beads were washed again thrice with ice-cold TBS as before. To elute bead-attached proteins, 50 µL of 0.2 M glycine (pH 2-3) were added to the beads and incubated for 10 min with strong mixing at RT. Then, 50 µL of 1 M Tris-HCl (pH 7-8) were added, alongside 20 µL of loading buffer (concentrated five times). Samples were analysed by WB as described above, and samples reserved from the protein extraction were used as inputs.

Structural In Silico Modelling of CaMKIIα Oxidation
The structure of the N-terminus (Nt) domain of CaMKIIα was taken from PDB [32] (code 6W4O) and the conformation of the CaM binding helix was completed with the structure of AlphaFold2 [33] model (AF-Q9UQM7-F1_model_v3) up to residue 311. We used MODELLER [34] to model the structure of the Nt domain with the CaM binding helix bound by CaM. First, we removed the restrictions of the helix with respect to the Nt domain; then, we forced the contacts with CaM derived from the complex of a helix peptide with CaM (structure taken from PDB with code 6XXX). For the sequence of mutant T286E we used the model of the complex of the native sequence as the template, while for mutant C280E/M291E we added the restrictions of α-helix to the Nt capping (residues 280-285). PSIPRED [35] was used for the prediction of secondary structure using the full length of CaMKIIα/CaMKII.

Study of CaMKIIα Oxidation by Immunofluorescence in Human Samples
Human cortical sections (5 µm) were treated with 0.15 M glycine and 10 mg/mL NaBH 4 (both in PBS) for 30 min. They were washed in deionized water, permeabilized with 0.3% Triton-100X in PBS for 1 h at 4 • C, and then blocked with 5% FBS + 1% BSA diluted in 0.1% Triton X-100 in PBS for 2 h. The following step was incubation o.n. at 4 • C with 1:200 rabbit anti-ox-CaMKII and 1:200 goat anti-CaMKIIα Abs. Samples were incubated the next day with Alexa Fluor 647 donkey anti-goat Ab (1:2000; Abcam), Alexa Fluor 488 donkey anti-rabbit Ab (1:2000; Invitrogen) and DAPI (1:1000) for 1 h at RT. They were then treated for 10 min with 4% Sudan Black (Sigma) in ethanol that was previously mixed o.n. at 4 • C, centrifuged at 3000× g for 20 min, filtered using common filter paper, washed with deionized water and mounted using Fluoromont. Digital images were taken with a Leica TCS SP8 confocal microscope and analysed with ImageJ software (version 1.53c).

Study of CREB Phosphorylation by Immunofluorescence in Murine Samples
Cortical brain samples were treated following the protocol described in Section 4.15. without using NaBH 4 . Then, samples were incubated o.n. at 4 • C with the primary Abs 1:400 rabbit anti-p-CREB (Cell Signaling, Danvers, MA, USA) and 1:100 mouse anti-Aβ (6E10, Covance, Princeton, NJ, USA). After washing the samples, they were incubated with Alexa Fluor 488 donkey anti-rabbit Ab (1:1000; Invitrogen), Alexa Fluor 555 donkey anti-mouse Ab (1:1000; Invitrogen) and DAPI for 1 h at RT. Digital images were taken with a Leica TCS SP5 confocal microscope and analysed with ImageJ software (version 1.53c).

Study of CREB Phosphorylation in Primary Hippocampal Neurons
Primary hippocampal neuronal culture was treated for 1 h with 100 nM oAβ 1-42 , 1 µM Kn-93 or a combination of both. Cells were fixed with 4% PFA in PBS for 15 min, permeabilized with 0.3% Triton-100X in PBS for 10 min and then blocked with 3% BSA in PBS for 30 min. Coverslips were incubated o.n. at 4 • C in a humid chamber with rabbit anti-pCREB (Ser133) (1:100; Cell Signaling) Ab. Coverslips were incubated with Alexa Fluor 488 donkey anti-rabbit Ab (1:2000; Invitrogen) and DAPI (1:1000) for 1 h at RT. Coverslips were mounted onto slides with Fluoromount. Digital images were taken with a Leica TCS SP5 Inverted confocal microscope and analysed with ImageJ software (version 1.53c).

Real Time PCR Analysis of Gene Transcription Due to CREB Phosphorylation
SH-SY5Y cells seeded in a 12-well plate (200,000 cells/well) were treated with either 1 µM Kn-93, 10 µM oAβ 1-42 or a combination of both, in Ham's F12 medium supplemented with 1% FBS, for 3, 6, or 24 h. Total RNA was isolated from cells using a NucleoSpin RNA isolation kit (Macherey-Nagel, Düren, Germany). SuperScript III reverse transcriptase system (Thermo Fisher Scientific) was used for complementary DNA synthesis. Quantitative real-time polymerase chain reaction was performed using SYBR Green (Applied Biosystems, Waltham, MA, USA) in a QuantStudio 12K Flex system (Applied Biosystems). GAPDH and HPRT1 were used as housekeeping genes for the quantification of relative gene expression using 2-∆∆Ct. Primers used for hCaMKIIα were commercially obtained from Sino Biological. We designed the other primers with the sequences indicated in Figure S2.

Statistical Analysis
Data are expressed as the mean ± SEM of the values from the number of experiments indicated in the corresponding figures. Student's t-test, one-way analysis of variance (ANOVA) or two-way analysis of variance (ANOVA) plus Bonferroni as the post-hoc test were used for statistical analyses. p > 0.05 is considered not significant (ns).