|Home | About | Journals | Submit | Contact Us | Français|
Hypoxic preconditioning (HP) is an evolutionarily-conserved mechanism that protects an organism against stress. The mitochondrial ATP-sensitive K+ channel (mKATP) plays an essential role in the protective signaling, but remains molecularly undefined. Several lines of evidence suggest that mKATP may arise from an inward rectifying K+ channel (Kir). The genetic model organism C. elegans exhibits HP and displays mKATP activity. Here, we investigate the tissue expression profile of the three C. elegans Kir genes and demonstrate that mutant strains where the irk genes have been deleted either individually or in combination can be protected by HP and exhibit robust mKATP channel activity in purified mitochondria. These data suggest that the mKATP in C. elegans does not arise from a Kir derived channel.
Brief periods of cell stress can activate endogenous protective mechanisms that reduce injury upon subsequent exposure to stress. Ischemic preconditioning (IPC) is one such example, whereby short periods of ischemia prior to the longer ischemic insult protect against ischemia-reperfusion (IR) injury. Since its discovery , IPC has been described in all species tested thus far including the genetic model organism C. elegans, where it has been termed hypoxic preconditioning (HP) . Mitochondrial K+ channels play an important role in protection, and the activation of the mitochondrial ATP-sensitive K+ channel (mKATP) is necessary and sufficient to protect against IR injury [3;4]. Since the mKATP is a critical component of IPC, the channel has been the subject of intense investigation and has been found in humans , rats , plants , amoeba , T. cruzi , and in C. elegans .
The molecular identity of mKATP remains elusive, but is thought to resemble the surface KATP channel, in which the pore forming subunit is a member of the inward rectifier potassium channel (Kir) family. Kir channels perform a diverse array of physiologic functions, and gene mutations have been associated with a variety of pathological conditions in humans [for review, see ]. The mammalian Kir gene family consists of fifteen isoforms divided into seven subfamilies (Kir1.x to Kir7.x). These subfamilies can be further classified into four functional groups, including classical, or constitutively active, G-protein gated, ATP-sensitive, and K+ transport channels. Surface KATP channels are composed of tissue specific combinations of inward rectifying K+ channel (Kir6.x) subunits and regulatory sulfonylurea receptor (SURx) subunits , and mKATP channels have been proposed to arise from some combination of these. In cardiac myocytes, the surface KATP and the mKATP resemble each other with respect to physiologic transport characteristics, but they can be distinguished by pharmacology. For example, diazoxide and 5-hydroxydecaonate (at appropriate concentrations) selectively activate and inhibit the cardiac mKATP, respectively . These reagents, however, are not completely specific for the mKATP, as they can also regulate surface channel activity in other cell types, leading to the idea that the pharmacologic dissimilarity between the surface and mitochondrial channels arises from some unique combination of Kir6.x and SURx, or novel splice variants . Nevertheless, despite the tools available and its critical role in the evolutionarily conserved endogenous mechanism of IPC, the mKATP has not been conclusively defined at the molecular level using pharmacologic techniques.
The use of genetic models has also failed to result in the molecular identification of the mKATP. A mouse knockout model of Kir6.2 was shown to be resistant to IPC , but this resistance was not related to the loss of the mKATP. This may be linked to the fact that mutations in both Kir6.2 and SUR1 have both been linked to familial persistent hyperinsulinemic hypoglycemia , coupled with knowledge that hyperglycemia can abrogate protection by IPC . Similarly, Kir6.1 and SUR2 knockout mice both display a phenotype indicative of Prinzmetal angina [18;19], complicating an analysis of its role in IPC. However, C. elegans lacks a vasculature and its genome codes for only three Kir family channels, irk-1, irk-2, and irk-3. Hence, we hypothesized that this genetic model, which can be protected by HP  and exhibits a mKATP activity  would be a useful reductionist model to assess the role of the Kir gene family in protective signaling.
Unlike mammals, the C. elegans family is limited to three genes, termed irk-1, -2, and -3. The reduced complexity of this family compared to that in humans would be expected to facilitate dissection of Kir function in worms. However, other than the fact that they are all most closely related to Kir2.x family members, with ~50% amino acid identities, there is very little known about them. Herein, we used fluorescent transgenes to probe the cellular expression pattern and protein localization of all three Kir genes (irk-1, -2, and -3) in C. elegans. Using worm genetic knockouts we determined if these genes were involved in HP or contributed to mKATP activity in isolated mitochondria. Our results suggest that the worm IRK channels have a limited distribution, which is inconsistent with a conserved function in mitochondria, and do not contribute to either HP or mKATP activity.
Worm strains were routinely propagated at 20°C on nematode growth media agar plates containing 5 μg/ml cholesterol and seeded with OP50 bacteria. Strains used in this study include Bristol-N2; fluorescent protein fusion strains (all in a pha-1(e2123ts)III, him-5(e1490)V genetic background) KWN73 rnyEx040 [pKT48 (Pirk-1::IRK-1::GFP); pCL1 (pha-1+)], KWN83 rnyEx043 [pKT54 (Pirk-3::GFP); pCL1 (pha-1+)], KWN88 rnyEx045 [pKT49 (Pirk-2::IRK-2::GFP); pCL1 (pha-1+)]; and irk gene deletion strains MT15934 irk-1(n4895)X, MT15935 irk-2(n4896)X, MT17369 irk-3(n5049)X, MT17360 irk-1(n4895)X, irk-2(n4896)X, irk-3(n5049)X (these mutants were generously provided by N. Ringstaad and R. Horvitz). The triple deletion was obtained by a conventional mating strategy with PCR genotyping used to confirm the presence of the mutant alleles in the final strain.
Transgenic C. elegans strains expressing the fluorescent irk gene fusions were placed onto 2% M9 agarose pads under anesthesia (1 mg/ml tetramisole) and examined using an Olympus FV1000 confocal microscope (available as part of the University of Rochester Confocal Core) under appropriate illumination.
Briefly, C. elegans were chronologically-synchronized and adult hermaphrodites were collected, washed and allowed to lay embryos for a period of two hours. The resulting adult progeny were subject to the following protocols as previously described . Simulated ischemia-reperfusion (IR) killing: C. elegans in M9 were incubated at 26°C for 20 hrs in an anoxic chamber (5%/95% H2/N2). Following the ischemic insult, C. elegans were returned to normoxia at 20°C for 24 hrs. Animals were then scored. Animals which lack movement (spontaneous or touch-evoked) were scored as dead. Hypoxic preconditioning (HP): C. elegans were incubated for 4 hrs in an anoxic chamber at 26°C. Following the stimulus, C. elegans were moved to normoxia at 20°C for 20 hours to allow for recovery. The preconditioned C. elegans were then subjected to IR killing as described above. For all experiments, seven replicates of at least thirty C. elegans were observed under paired IR and HP protocols for each of the mutant strains.
Mitochondria were isolated from C. elegans as previously described . The mitochondrial suspension was placed in a chilled stirring cuvette and incubated in the dark with 20 μM BTC-AM (Invitrogen) and 0.05% Pluronic F-127 for 10 min at room temperature. Mitochondria were washed to remove excess BTC-AM by centrifugation at 7000 × g for 10 min. The pellet was resuspended in MIM (220 mM mannitol, 70 mM sucrose, 5 mM MOPS, 2 mM EGTA, 0.4% BSA, pH 7.4, filtered), stored on ice in the dark and used within 1.5 hours of isolation. Protein was determined by the Folin-phenol method .
mKATP activity was determined by monitoring the uptake of Tl+, a surrogate for K+, into mitochondria using a Varian Cary Eclipse spectrofluorometer as previously described . Briefly, ~0.8 mg/ml BTC-AM-loaded mitochondria were added to a rapidly stirred cuvette containing 2 ml of Tl+ assay buffer (2 mM pyruvate, 2 mM malate, 195 mM mannitol, 10 mM HEPES, 2 mM MgCl2, 2 mM Na2HPO4, and 1 μg/ml oligomycin, pH 7.2 at 25°C). Compounds were present in the buffer and incubated with mitochondria for 30 seconds. Tl2SO4 (2 mM) was added via a syringe port and the change in fluorescence was monitored (λex 488 nm, λem 525 nm) and normalized to an initial baseline recording 10 seconds.
Chemicals were of the highest grade available from Sigma (St. Louis, MO) unless otherwise specified. BTC-AM was from Invitrogen (Carlsbad, CA). Atpenin A5 was from Axxora LLC (San Diego, CA),
In isolated mitochondrial studies, statistical significance (P<0.05) between multiple groups was determined using analysis of variance (ANOVA). In whole worm studies, significance (P<0.05) was determined using a paired Student’s t-test. Data presented are mean ± SEM.
C. elegans hermaphrodites are comprised of ~1000 somatic cells that share many basic cell biologic processes with mammals. In order to define the cells and cell types that express the irk genes, each of the three irk promoters was amplified using genomic PCR and fused via PCR sewing to the coding region of GFP. These PCR products were used to create transgenic nematode strains where the fluorescent cells demarked the limits of irk gene expression (transcriptional reporters Pirk-x::GFP). As a complementary approach, the entire irk coding sequence was amplified together with its cognate promoter and fused in-frame to GFP (translational reporters Pirk-x::IRK-x::GFP). Translational reporters give rise to fluorescent IRK fusion proteins that can be used to determine intracellular protein localization. It should be noted when examining worms expressing these constructs that promoter-driven transgenes may not completely recapitulate endogenous expression patterns, either due to mosaicism (where individual worms may not exhibit the full repertoire of cells due to loss of the transgene during mitosis) or because sequences in the genomic region such as the 3′ UTR contribute to proper expression.
Figure 1 shows representative images of transgenic strains expressing fluorescent reporters for each of the irk genes. For the irk-1 and irk-2 genes, their respective transcriptional and translational fusions were expressed in a similar repertoire of cells (data not shown), and hence only the translational fusions are shown. The irk-1 promoter drove GFP expression in multiple neurons, including the hermaphrodite specific neuron (HSN), sensory phasmid neurons in the tail, and unidentified neurons in both the head and tail of male worms (Figure 1A–D). Intestinal and gonadal cells were labeled as well (Figure 1E–F). As would be predicted for all of the IRK proteins based upon homology with the mammalian Kir channels, clear plasma-membrane targeting of the GFP-tagged IRK-1 protein can be observed in the intestine (Figure 1C and 1E). This does not preclude dual targeting of the protein to the mitochondria, however, and it should be noted that enforced expression of fluorescent fusion proteins in worms can lead to targeting artifacts. However, the irk-1 cDNA expressed in Xenopus oocytes gave rise to a plasma membrane K+ conductance (N. Ringstad, personal communication), suggesting that IRK-1 is a bona fide plasma membrane channel. The fusion protein was likely active, as well, since strains expressing it were viable but were not healthy, and exhibited signs of cellular degeneration over time. For example, the neuronal processes in adults assumed a tortuous morphology and the fusion proteins appeared to aggregate over time (Figure 1A). This can occur due to overexpression or misregulation of a “leak” conductance type channel that is normally only active under select conditions.
As for IRK-1, IRK-2 is also expected by be targeted to the plasma membrane. The most notable cells labeled by irk-2 promoter expression were neurons of the anterior ganglia, the pharyngeal intestinal valve cell, and the posterior intestine, where again clear basolateral plasma membrane targeting can be observed (Figure 1G–L). In adults, the spermatheca and the uterus were prominently labeled (Figure 1K), and the IRK-2 fusion protein resided in the plasma membrane.
The irk-3 translational fusion was not tolerated by the worms, even at low doses. Hence, a transcriptional fusion was used to assess the extent of irk-3 gene expression. Irk-3 promoter activity was limited to neurons, but it was broadly expressed throughout multiple neuronal subtypes, including sensory neurons and the HSN (Figure 1M–O).
It is interesting that irk gene expression was notably absent in muscle and hypodermal cells. These cell types are amongst the largest in C. elegans and would be expected to contribute greatly to their total mitochondrial content. A channel that is localized to the mitochondria, even in part, and contributes to mitochondrial protection from insult might be predicted to be expressed more ubiquitously than observed here for the irk genes.
Since HP is widely conserved across species and is known to protect C. elegans from hypoxic death , and worms mitochondria exhibit robust mKATP channel activity , we sought to determine whether the C. elegans mKATP channel is comprised of a Kir family gene product. Mammals contain fifteen Kir genes, while C. elegans only contains three genes in the Kir family, termed irk-1, irk-2, and irk-3. As is often the case with large gene families, it is difficult to identify mammal-worm orthologs, and the worm genes are more closely related to one another than to any of the mammalian isoforms. The mammalian gene product most closely related to the worm irk family is Kir2.1. However, it should be noted that KATP channels open spontaneously in the absence of ATP and can be broadly classified with the classical Kir2.x channels which have constitutive activity.
C. elegans loss-of-function mutants in the irk-1, irk-2, and irk-3 genes are viable, and due to the absence of tissues that are known to depend on KATP function in mammals (e.g. pancreatic beta-cells) in the worm, we hypothesized that they would be less likely to present confounding issues. First, we sought to determine if the irk gene products were required for HP. The worm HP model is a model of delayed preconditioning where a short 4 hr hypoxic exposure, followed by 20 hrs of normoxic recovery, induces resistance to a subsequent 24 hrs hypoxic killing. As shown in Figure 2, HP significantly protected the wild-type (WT) strain from hypoxic stress, as previously demonstrated [2;21]. Unexpectedly, all of the irk deletion strains were also protected by HP to the same extent as WT, as was a triple deletion strain containing combined irk-1, irk-2, and irk-3 mutant alleles (Figure 2). However, the triple deletion strain exhibited somewhat less variability in its basal sensitivity to hypoxia than the WT and, though statistically insignificant, appeared slightly more resistant to hypoxia on average (WT, 60% death; triple, 50% death; p-value 0.30). In conclusion, these results suggest that HP in C. elegans is independent of a Kir family K+ channel.
We next sought to determine whether C. elegans mKATP activity was affected by the loss of irk gene function in the deletion mutants. The mKATP contributes to K+ transport across the mitochondrial inner membrane, and channel activity can by monitored by fluorescent measurements of Tl+ fluxes (as a surrogate for K+) in isolated mitochondria using the dye BTC-AM [20;21]. Using this technique, we measured changes in mitochondrial Tl+ uptake in response to a conserved repertoire of activators and inhibitors that together define a pharmacologic signature for mKATP activity. We then compared this signature in WT worms to that in the irk-1/-2/-3 triple deletion mutant to assess the contribution of worm irk genes to mKATP activity.
In WT worms we observed mitochondrial Tl+ flux that was inhibited by ATP and activated by both diazoxide and atpenin A5 (Figure 3). This activation by channel openers could be inhibited by mKATP blockers 5-hydroxydecaonoate and glyburide (Figure 3), consistent with mKATP activity in isolated C. elegans as well as mammalian mitochondria . Mitochondria from the triple irk deletion strain exhibited a nearly identical profile, providing further evidence that mKATP is not derived from a Kir protein in C. elegans.
The mKATP is central to IPC-mediated protection yet its molecular identity has remained elusive. The simplest hypothesis put forth thus far has been that the protein composition of the mKATP resembles that of the defined surface KATP, which is comprised of an octamer of four Kir channel subunits and four SUR accessory subunits. However, our results demonstrate that neither HP nor mKATP requires Kir gene function in the nematode C. elegans.
One interesting outcome of the current data is the possibility that Kir genes are also not involved in the mKATP in mammals. Significant effort has been directed toward assessing a role for putative KATP components (e.g. Kir6.2, Kir6.1 and SUR) in mammalian mitochondria, with ambiguous results. These efforts can be broadly categorized into three areas. The first of these relies upon antibody detection, the second on genetic elimination of candidates, and the third on pharmacologic manipulation of channel activity.
Approaches based primarily on antibody labeling concluded that either Kir6.2 , Kir6.1 [22–24], both [25–27] or neither  were present in mitochondria, perhaps due to variability in reagent quality. In this regard, two commercially available anit-Kir6.1 antibodies were shown to recognize off-target mitochondrial proteins, NADH-dehydrogenase flavoprotein 1 and isocitrate dehydrogenase . Using custom antibodies, a novel splice variant of SUR2 has been identified in cardiac mitochondria , but its relevance to mKATP is as of yet unclear.
To attempt to circumvent this type of confusion, genetic loss-of-function mouse models have been developed for Kir6.1 , Kir6.2 , SUR1 , and SUR2 . However, these models exhibited either incomplete loss of gene activity , or confounding effects (e.g. elevated blood glucose) that prevented unambiguous interpretation of their ability to be effectively protected by IPC [16–19]. Despite this, a technique utilizing flavoprotein fluorescence has been employed as an indirect measurement of mKATP activity in genetic deletion or dominant negative gene transfer models, and neither Kir6.1 nor Kir6.2 has been found to play an important role in mKATP opening-induced flavoprotein fluorescence [15;33]. This result is in general agreement with our finding that none of the Kir genes plays a role in mKATP activity in worms. Furthermore, the mKATP has been described in plants recently, where the single Kir gene was hypothesized to be unlikely to contribute to the formation of the mKATP . Clearly, genetic deletion of all 15 Kir isoforms in mammals is not practical, but these results do suggest that the search for the mKATP should explore other non-Kir candidates.
Pharmacologic approaches to studying mKATP have been similarly ineffective at channel identification. In general, compounds which open the surface KATP activate the mKATP and cause cardioprotection. Likewise, inhibitors of the surface KATP block the mKATP. This cross-reactivity suggests some degree of conservation between these two channels, but makes it difficult to study the role of either one in isolation. This fact along with other shortcomings has caused some to question whether a mKATP channel contributes to cardioprotection or exists at all . However, there are a few pharmacologic tools that have used to dissect mKATP from the surface KATP. At low concentrations in cardiac myocytes, diazoxide and 5-HD are specific mKATP modulators and do not regulate the surface channel . However, even in myocytes at higher concentrations they have additional effects that might be predicted to impact IPC independent of their effect on mKATP [36;37]. These alternate targets may actually provide some insight into the molecular nature of the mKATP.
For example, the specific activator of the mKATP, diazoxide, also inhibits complex II of the mitochondrial respiratory chain . This characteristic is not unique to diazoxide, as the same result was observed for malonate  and atpenin A5  amongst other complex II inhibitors . Yet another study demonstrated that a purified mitochondrial inner membrane fraction displayed K+ channel activity that was sensitive to mKATP activators and blockers but also responded to complex II inhibitors . Hence, a moonlighting role for complex II in mitochondrial K+ transport has been hypothesized.
While the data presented herein demonstrate that a Kir isoform is not involved in the composition of the mKATP, they do not rule out the participation of a SUR subunit. In this regard, mKATP activity can be inhibited with the sulfonylurea glyburide (glybenclamide) even in the irk triple mutant (Figure 3). It is possible that a SUR might associate with an alternate K+-transporting subunit to give rise to mKATP activity. It is also possible, as discussed for diazoxide, that glyburide may have an alternative target [41;42]. In this respect, gyburide has been shown to interact with mitochondrial ADP/ATP carriers . Further work is undoubtedly necessary to elucidate the role of a SUR subunit in the composition of the mKATP.
Another interesting candidate arises from recent work examining the regulation of mKATP by redox status . While a general trend of channel inhibition was observed with reducing agents of increasing redox potential, it was observed NADPH did not conform to this trend, i.e., despite identical redox potentials, NADPH was far more inhibitory to channel function than NADH. This suggests that NADPH is able to inhibit mKATP via a non-redox mechanism. Notably, the beta subunit of the “Shaker” family Kv channels has been shown to contain an NADPH binding side, coupled to an aldoketoreductase enzymatic activity [44;45]. In addition, single channel patch-clamp experiments from plant mitochondria are consistent with a Shaker like conductance, though the electro-physiologic characteristics of the single plant Kir gene product are unknown . This raises the intriguing possibility that an NADPH binding Kv-beta subunit may be involved in mKATP assembly and/or regulation. Similarly intriguing perhaps is the observation that Kv1.3 forms a mitochondrial channel in lymphocytes that regulates apoptosis [46;47].
The quest for the molecular identity of mKATP channel remains. Antibody based approaches, pharmacologic regulatory strategies and mammalian genetic models have led to ambiguous conclusions and suffer from confounding effects. Furthermore, compensatory mechanisms involving other Kir subunits may rescue the mKATP activity in knockout animals. The data presented herein demonstrate effective HP and normal mKATP activity in C. elegans that do not have any of the three inwardly rectifying K+ channel genes. Furthermore, the use of a strain where all the Kir family genes in C. elegans have been knocked out together rules out the possibility of compensatory mechanisms involving another Kir isoform. These results suggest that the evolutionarily conserved phenomena of the mKATP and IPC may not center on a Kir component and should drive investigation of alternate targets.
We thank Robert Horvitz and Niels Ringstad (Howard Hughes Medical Institute, Cambridge, MA and New York University, New York, NY) for the irk deletion strains. This work was supported in part by an American Heart Association, Founder Affiliate Postdoctoral Fellowship award 11POST7290028 (to APW), and by US National Institutes of Health grants HL071158 (to PSB), GM087483 (to PSB & KWN).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.