EVALUACIÓN DEL HUMEDAL ARTIFICIAL

To determine if lobster HCN and Kv4 channels, which respectively mediate Ih and IA,

were SUMOylated in vivo, we performed immunoprecipitation (IP) experiments with lobster

CNS protein lysates followed by western blotting. IP experiments were carried out using custom affinity-purified antibodies against the lobster HCN channel (anti-HCN), the lobster Kv4 channel (anti-ORF-J) and IgG (negative control). Western blots containing the resulting IP products were probed with a commercially available affinity-purified antibody against SUMO (anti-SUMO), and either anti-HCN (Fig 4.2A) or anti-ORF-J (Fig 4.2B). Both HCN and Kv4 channels appeared

to be SUMOylated in vivo. The anti-HCN and anti-SUMO antibodies recognized multiple

matching bands in the anti-HCN but not the IgG IP product. The bands were in the correct size range for post-translationally modified HCN channel alternately-spliced isoforms (Ouyang, Goeritz et al. 2007). The anti-ORF-J and anti-SUMO antibodies recognized a single, appropriately sized, matching band in the anti-ORF-J but not the IgG IP product. The anti- SUMO antibody recognized additional bands not recognized by the channel antibodies in both

the anti-HCN and anti-ORF-J but not the IgG IP products. These may represent SUMOylated proteins that non-covalently interact with the channels.

4.4.4 SUMOylation is necessary for activity-dependent changes in LP Ih and IA

We next determined if Ih and IA were affected by anacardic acid, a cell permeable

SUMOylation blocker (Fukuda, Ito et al. 2009). Either Ih or IA was measured, then 100µM

anacardic acid was or was not added to the superfusate and a 1hr 0%ΔDC voltage protocol was applied followed by TEVC to re-measure the current. No significant changes were observed in

the absence of anacardic acid (Wilcoxon test, p>0.1, n=3 each for Ih and IA). In the presence of

anacardic acid, the mean LP Ih Gmax significantly decreased by 14±4.5% and the mean peak IA

significantly increased by 9.7±2.8% (Fig 4.3A). These data suggest that dynamic SUMOylation

contributes to the maintenance of LP Ih and IA. We asked if anacardic acid could also block

activity-dependence. Experiments similar to those diagrammed in Fig 4.1A were repeated except that anacardic acid was added to the superfusate for 1hr prior to the addition of TTX. The data indicated that blocking SUMOylation with anacardic acid abolished the DA-enabled activity-

dependence of LP Ih (Fig 4.3B) and the activity-dependence of LP IA in the absence of DA (Fig

4.3C). These data indicated that SUMOylation was necessary for activity-dependent changes in LP Ih and IA.

4.4.5 Enhanced SUMO availability converts bi-directional into one-way DA-enabled LP Ih activity-dependence

If dynamic SUMOylation regulates LP Ih Gmax, then enhancing SUMOylation should

have the opposite effect of blocking SUMOylation. Furthermore, if DA enables activity-

dependent SUMOylation, then enhancing SUMOylation should disrupt the permissive effect of tonic 5nM DA. We tested these hypotheses. The enzyme Ubc9 covalently links SUMO to a

lysine on a target protein (Desterro, Thomson et al. 1997). Ubc9 can either recognize a SUMO consensus sequence on a target protein and conjugate SUMO to the consensus sequence, or Ubc9 can be directed to a target by an E3 ligase or a SUMO Interaction Motif (SIM), in which case SUMOylation does not occur at a consensus sequence (Flotho and Melchior 2013).

Approximately 63% of all SUMOylation occurs at consensus sequences, though this can change in a state-dependent manner (Hendriks and Vertegaal 2016). Simply increasing the concentration of SUMO and/or Ubc9 can enhance SUMOylation at consensus sequences (Dai, Kolic et al. 2009). We used this method to increase mouse HCN2 channel SUMOylation at an identified consensus sequence in a heterologous expression system, which resulted in augmentation of

channel surface expression and Ih Gmax (Parker, Welch et al. 2016). Here we used the same logic

to increase SUMOylation at consensus sequences in the LP neuron to test if/how this affected Ih

Gmax and activity-dependent regulation of Ih.

We cloned lobster SUMO cDNA (GenBank accession: MF770707) and used standard techniques to generate an N-terminal Tat-tagged SUMO peptide (Supp Fig 4.1). The activated form of the peptide that ends in a C-terminal di-glycine was created; it required no further proteolytic processing before addition to a target (See Supp Fig 4.1)(Mukhopadhyay and Dasso 2007). Peptides tagged with an HIV-Tat sequence are rapidly taken up by cells (Fawell, Seery et al. 1994). After confirming that the activated Tat-SUMO peptide was successfully taken up by

crustacean neurons (Supp Fig 4.1), we tested its effect on LP Ih Gmax.

The three standard voltage protocols were used in experiments as diagramed in Fig 4.4A. A 30min wash-in of activated Tat-SUMO had the opposite effect of blocking SUMOylation; it

produced a stable and significant 11±1.5% increase in the mean LP Ih Gmax in the absence of DA

The mean τ for the increase was 7.1min. The same result was obtained when the Tat-SUMO peptide was applied during a 30min -100%ΔDC or 0%ΔDC voltage protocol (Fig 4.4B), indicating that an activity-independent process transduced enhanced SUMO availability into an

increase in LP Ih Gmax. Note that DA-enabled event 1 produced an activity-independent, ~12%

increase in LP Ih with a time constant of 9.9min (Krenz, Rodgers et al. 2015). Thus, enhancing

SUMO availability in LP mimicked DA-enabled event 1. Enhanced SUMO availability also occluded DA-enabled event 1. A 30min activated Tat-SUMO wash-in was followed by

application of the -100%ΔDC protocol in the presence or absence of 5nM DA. The data showed that application of the -100%ΔDC protocol in the presence of DA produced no further change in LP Ih (Fig 4.4C).

Enhanced activated SUMO availability mimics and occludes DA-enabled event 1. An increase in activated SUMO ultimately translates into enhanced target protein SUMOylation. Lobster HCN channels are targets of SUMOylation (Fig 4.2A). HCN channel SUMOylation can augment HCN channel surface expression (Parker, Welch et al. 2016). Together, these data suggest that DA-enabled event 1 may represent enhanced HCN channel SUMOylation.

Using the same experimental protocol (Fig 4.4A), we also examined whether or not enhanced SUMO availability blocked the DA-enabled, calcineurin-mediated, activity-dependent

decrease in LP Ih (event 2). Enhanced SUMO availability did not obviously alter event 2 (Fig

4.4C). In the presence of tonic 5nM DA and limitless activated SUMO, the +50%ΔDC protocol

still produced a significant average 9.5±1.8% decrease in LP Ih Gmax. This decrease was not

observed in the absence of tonic DA. Thus, enhanced SUMO availability did not block the