Heme Oxygenase: Localization
The heme oxygenases are ubiquitiously expressed key enzymes serving an essential role in oxidative heme degradation. Two main isoforms are expressed in mammals: HO-1 and HO-2. While HO-2 is constitutively expressed under basic requirements in the majority of human tissues, HO-1 represents the inducible HO variant whose expression is highly inducible by its substrate heme and different physical and chemical stressors 2-4. The expression of HO-1 is low under physiological conditions in most tissues, except for those implicated in heme metabolism such as liver, red blood cells, and bone marrow with highest levels found in spleen. In contrast, constitutive expression of HO-2 can predominantly be detected in brain, testes, spleen, and liver as well as the cardiovascular and nervous system 67, 77, 179. HO-1 and HO-2 have originally been characterized as being associated with membranes of the smooth ER. Both HO isozymes are highly homologous and are imbedded in the microsomal membrane via a similar C-terminal TMS 67. Intriguing reseach clearly revealed an association of HO-1 with extra-ER compartments, including the cytosol 5, 6, mitochondria 7, 8, plasma membrane caveolae 9, chloroplasts 10-13, extracellular space 14, and the nucleus 15. Anchorage of HO-1 to the ER, caveolae and mitochondria obviously occurs through its TMS 21. In the ER, caveolae and mitochondria, TMS-anchored HO-1 appears to co-localize with cytochrome p450 reductase (CPR) and biliverdin reductase (BVR), implying heme degradation as being its primary function 21. Nuclear HO-1 was found as being catalytically inactive, but it might function as a transcriptional regulator modulating its own expression and that of the transcription factor AP-1 15, 191. In these studies, it was shown that the loss of the TMS induced by different triggers is required for the nuclear translocation of HO-1. Mitochondrial HO-1 has been found as being associated with the mitochondrial inner membrane even though it lacks a regular mitochondrial targeting sequence 7. It has been suggested that mitochondrial-associated membranes (MAM) might be implicated in the trafficking of HO-1 between mitochondria and the ER 21, 192. In the plasma membrane, HO-1 could initially be detected in caveolae of pulmonary artery endothelial cells (ECs) from rats 9 where it interacts with the caveola-specific membrane proteins caveolin 1 (Cav-1) and Cav-2, respectively 9, 193, 194. HO-1 is supposed to translocate to caveolae and the plasma membrane via the ER and Golgi apparatus, where it contributes to the detoxification of heme. The interaction of HO-1 with Cav-1 was noted to reduce the activity of HO-1, indicative for a negative regulatory role of Cav-1 9. Plants express several HOs but these often lack the C-terminal TMS found in mammalian HOs. In higher plants, HO-1 has been found predominantly in the stroma of chloroplasts 12. Plant HOs can be encoded by both, the nuclear and the plastidic genome. As already mentioned, a nuclear location of HMOX genes can be detected, e.g., in higher plants, in the green alga C. reinhardtii, and in C. caldarium while the corresponding genes in P. purpurea and R. violacea are of chloroplastic origin. The model plant A. thaliana expresses four plastidic HOs (AtHO-1 to -4) that are nuclear encoded and translated in the cytosol. These nuclear encoded HOs contain N-terminal chloroplast translocation sequences that efficiently target them to chloroplasts. In contrast, the pbsA gene was found in the plastidic genome of most marine red algal species and cryptophytes 195. It is worth mentioning that, in most of the PbsA proteins from red algae, a putative transmembrane domain was predicted at the C-terminus. Plastidic genome analysis by Cho and colleagues also revealed that pbsA in C. paradoxa (Glaucophyta) was located to the nucleus, and that the expressed PbsA protein exhibits the critical N-terminal transit peptide 195.
Besides its intracellular location, HO-1 can be found in various extracellular compartments including body fluids. HO-1 was identified not only in human milk 196 but also the serum and plasma as well as the cerebrospinal fluid of healthy and diseased human individuals (summarized by Vanella et al., 2016) 17. Extracellularly located HO-1 might function as a potential biomarker in disease 18, 19 or as an extracellular receptor ligand 20. The molecular nature of the HO-1 release is, at least in part, enigmatic as HO-1 lacks the consensus signal for secretion. Consequently, HO-1 cannot be released from cells by the conventional secretory pathways. As outlined before, one mechanism underlying its export is represented by an active secretory process with the participation of caveolae, a specialized version of membrane-associated lipid rafts. Additional mechanisms of HO-1 release may involve a number of non-canonical pathways including secretory lysosomes, microvesicles such as ecto- or exosomes, release in free form as well as necrotic mechanisms resulting in a passive leakage out of cells 197. In this context it is worth highlighting that, in contrast to HO-1, HO-2 could not be detected in extracellular compartments implying that the export of HO-1 is unlikely to be caused by a passive leakage of the enzyme.