Molecular Bases of Variegate Porphyria

Bad Genes –Inheritance of Variegate Porphyria and its Impact on Heme Biosynthesis

Variegate porphyria is an autosomal dominant disorder that results from mutations in the PPOX gene located on chromosome 1q22-23 (Barbaro, 2013). Over 130 of such mutations have been identified (Qin, 2011), all of which reduce PPOX activity by 50% (Wang, 2013). In South Africa, where the disease is most common, >95% of cases involve the so-called “founder mutation” which was brought over by Dutch Settlers in 1688 (Dooley, 2007). This R59W mutation reduces the activity of PPOX by 50%, resulting in the accumulation of heme precursors within the body (Wang, 2013).

Interestingly despite being an autosomal disease, VP shows only 40% penetrance (Hift, 2012), appears more often in women than men (Singal, 1993), and does not often manifest until puberty (Siegesmund, 2010). In an attempt to explain these interesting aspects of VP, du Plessis et al. studied the transcription of the ALAS-1 gene in 26 patients with the R59W form of VP. Since past work on the PPOX promoter region yielded no correlation to the variance in expression, du Plessis et al. decided to look at ALAS-1 as ALAS-1catalyzes the rate-determining step of heme biosynthesis (du Plessis, 2009). After sequencing PCR fragments from the VP patients, they found two variants in the 5’-regulatory region of ALAS-1: -853C/T and -1253T/A; computational studies revealed -853 is located immediately 5’ to a half-palindromic estrogen response element.

In order to assess the role of estrogen on ALAS-1 transcription, the authors constructed three plasmids containing the PPOX promoter and the gene that encodes luciferase. Plasmid 1 had -853C and -1253T, plasmid 2 -853T and -1253A, and plasmid 3 -835C and -1253T. These were then used to transform HepG2 cells and the impact of estrogen treatment on luciferase activity was assessed. Results (Figure 1) indicate the HepG2 cells expressing plasmid 3 showed increased luciferase activity upon treatment with estrogen. In addition, HepG2 cells with plasmid 2 had less luciferase activity than those transformed with plasmid 1. Collectively, this indicates that variations in the ALAS-1 promoter region can impact the rate of ALAS-1 transcription antagonistically. Depending on which variants an individual has will dictate the likelihood of the disease manifesting itself. In addition, the sensitivity of ALAS-1 transcription to estrogen explains the tendency of VP to appear in women, as they produce more estrogen in men, and for symptoms to appear at puberty, when hormones are produced at high levels.

Figure 1. Histogram showing luciferase activity in HepG2 cells transformed with one of three plasmids. From top to bottom: -853T and -1253T, -853C and -1253A, -853C and -1253T. Source: Du Plessis, 2009
Figure 1. Histogram showing luciferase activity in HepG2 cells transformed with one of three plasmids. From top to bottom: -853T and -1253T, -853C and -1253A, -853C and -1253T. Source: Du Plessis, 2009

Role of Protoporphyrinogen Oxidase in Variegate Porphyria

Protoporphyrinogen oxidase is an inner mitochondrial membrane protein that requires both FAD and molecular O2 to function (Qin, 2011). The reaction catalyzed by PPOX, the conversion of protoporphyrinogen IX to protoporphyrin IX, is proposed to occur in three steps, each of which involve a two-electron-transfer from the substrate to a molecule of O2 via FAD in order to form H2O2 (Qin, 2011). A proposed mechanism for this process has been elucidated (Scheme 1) (Koch, 2004).

Scheme 1. Proposed PPOX mechanism. Source: Koch, 2004.
Scheme 1. Proposed PPOX mechanism. Source: Koch, 2004.

In 2011, Qin et al. reported the crystal structure of human PPOX in complex with FAD and the inhibitor acifluorfen at a resolution of 1.9 A (Figure 2).

Figure 2. Crystal structure of PPOX bound to FAD and acifluorfen. Source: Qin, 2011.
Figure 2. Crystal structure of PPOX bound to FAD and acifluorfen. Source: Qin, 2011.

This crystal structure shows the three PPOX domains: FAD-binding, substrate-binding, and membrane-binding. The FAD-domain contains 13 β-strands and 8 α-helices, and FAD is recognized by the canonical βαβ-super secondary structure (Qin, 2011). Key interactions include multiple hydrogen bonds between the ribose, adenine, and pyrophosphate moieties and specific residues, plus stacking between the isoalloxazine ring and Pro58 (Qin, 2011). The substrate domain consists of 5 β-strands and 3 α-helices (Qin, 2011). Using in silico docking, the authors were able to study the binding of protoporphyrinogen IX to the crystal structure of PPOX. This revealed a number of important interactions including hydrogen bonds to stabilize the proprionic acid carboxyl groups and hydrophobic interactions accommodating the four pyrrole rings (Figure 3). The docking study also indicates the methylene unit believed to transfer electrons from protoporphyrinogen IX to FAD is located 3.3 A away from the N5 electron acceptor of FAD, a reasonable distance for electron movement (Qin, 2011).

Figure 3. Crystal structure showing the binding of protoporphyrinogen IX to the active site of PPOX. Source: Qin, 2011.
Figure 3. Crystal structure showing the binding of protoporphyrinogen IX to the active site of PPOX. Source: Qin, 2011.

Qin et al. also expressed the R59W mutant in E. coli in order to study the mutation’s effect on PPOX. Since Arg59 is located in the FAD-domain, the decreased activity of PPOX in VP is likely due to a decrease in binding affinity for the cofactor. Replacement of the positively charged Arg residue with a bulky, hydrophobic Trp likely causes electrostatic conflict due to repulsion between the hydrophilic binding site and the indolyl ring of tryptophan (Qin, 2011). This forces the Trp side chain away from the hydrophilic residues, resulting in conformational changes of neighboring amino acids that disrupt their ability to interact with and bind FAD, decreasing its affinity for PPOX (Figure 4).

Figure 4. Comparison of the WT PPOX FAD binding site to that in the PPOX mutated with the "founder mutation". Source: Qin, 2011.
Figure 4. Comparison of the WT PPOX FAD binding site to that in the PPOX mutated with the “founder mutation”. Source: Qin, 2011.

Lastly, the membrane-binding domain contains 8 α-helices and 2 β-strands (Qin, 2011). It is believed that the 8 helices serve as the membrane anchor (Koch, 2004). Interestingly, up until last year, it was believed that PPOX faced the intermembrane space of the mitochondria. However, proteomic mapping of mitochondria in living cells by Rhee et al. demonstrate that it actually faces the mitochondrial matrix (Figure 5). This new model of PPOX binding has big implications for its interactions with ferrochelatase and corproporhyrinogen oxidase, both of which are believed to form physiological complexes with PPOX to form temporary channels by which heme precursors travel between active sites (Koch, 2004).

Figure 5. Comparison of the new (left) and old (right) models of PPOX binding to the intermitochondrial membrane. Source: Rhee, 2013.
Figure 5. Comparison of the new (left) and old (right) models of PPOX binding to the intermitochondrial membrane. Source: Rhee, 2013.

Global Impacts of Variegate Porphyria on the Body

The decreased activity of PPOX causes the accumulation of both porphyrins and porphyrin precursors in the body, resulting in the manifestation of both hepatic and erythropoietic symptoms. Erythropoietic traits of VP are believed to arise from oxidative damage induced by porphyrin accumulation in fat. Due to the conjugation present in these molecules, they are able to absorb UV radiation and produce reactive oxygen species (ROS) that can then induce lipid peroxidation that eventually manifests as bullae, hyperpigmentation, and the like (Sassa, 2006).

Even though ROS are believed to play a role in the pathogenesis of VP, little has been done on studying the state of oxidative stress in VP patients.  In order to better understand the impact of PPOX on the redox state of cells, Ferrer et al. examined the state of oxidative stress in the lymphocytes of 12 women with VP, compared to those from 12 healthy women, since these cells tend to serve as a good model for oxidative stress in the body as a whole (Ferrer, 2010).

First, the authors looked at the rate of H2O2 production to assess the ability of the cells to form ROS. While they found no significant difference between the production of H2O2 in VP cells versus the control under basal conditions, stimulation of the cells with phorbol meristate acetate (PMA) induced a higher amount of ROS production in the VP cells (Figure 6) (Ferrer, 2010). In order to determine the source of the ROS, the effects of three inhibitors were tested: myxothiazol, an inhibitor of complex III, rotenone, an inhibitor of complex I, and allopurinol, an inhibitor of xanthine oxidase. Only myxothiazol returned activated cells to the basal level of H2O2 production, indicating this component of the electron transport chain is responsible for most of the ROS production in VP (Figure 6). Since many of the components of this complex require heme, it is not too surprising that a disease involving disruptions in heme biosynthesis should show increased ROS formation of this site of the electron transport chain (Ferrer, 2010).

Figure 6. Comparison of H2O2 production in lymphocytes from healthy (white) and VP patients (black). From left to right: basal conditions, PMA stimulation, myxothiazol, retanone, and allopurinol. Source, Ferrer, 2010.
Figure 6. Comparison of H2O2 production in lymphocytes from healthy (white) and VP patients (black). From left to right: basal conditions, PMA stimulation, myxothiazol, retanone, and allopurinol. Source, Ferrer, 2010.

After demonstrating the susceptibility of porphyric cells to make ROS, the authors checked the expression of three proteins known to be involved in regulating ROS formation at the electron transport chain: UCP-3, Bcl-2, and SIRT-3. Results indicate that UCP-3, Bcl-2, SIRT-3, and PPOX are all expressed at significantly lower levels in VP relative to the control. Taking it a step further, the authors connected this imbalance of increased susceptibility to form ROS and decreased antioxidant defenses to the rate of oxidative DNA damage (Ferrer, 2010). As was seen with H2O2 production, basal conditions did not involve a significant difference between the healthy and control, but after exposure with H2O2, VP cells showed much more oxidative damage.

In summary, Ferrer et al. demonstrate that lymphocytes of VP women are in a state of oxidative stress. While ROS formation is not exacerbated under basal conditions, after stimulation, the cells produce much more H2O2 than the healthy control. In addition, the decreased expression of Bcl-2, UCP-3, and SIRT-3­ demonstrate an impaired ability to combat these molecules, resulting in increased oxidative DNA damage. This provides some insight into the mechanism by which VP can produce erythropoietic symptoms, as exposure of porphyrins to UV light is likely a stimulus for the production of ROS, resulting in oxidative damage near the skin where UV radiation penetrates best.

In addition, VP patients can also present hepatic symptoms. Though less understood, it is believed these result from excessive amounts of aminolevulinate (ALA) and/or porphobilinogen in the body. Possible mechanisms include: the ability of ALA to bind and damage GABA receptors, alterations in tryptophan metabolism, and/or a deficiency of heme in nerve cells (Meyer, 1998). Recently, ALA has also been linked to the destruction of myelin by the oxidation of neuronal proteins and lipids (Felitsyn, 2009).

First, Felitsyn et al. examined the impact of ALA on myelin formation. Control cells produced abundant amounts of myelin after 10 days while cells treated with ALA showed markedly reduced levels. Next, they looked at the impact of ALA on preformed myelin, and found that after 12 days of ALA exposure, neural cells showed destruction of the previously made myelin. Thus, ALA appears to both destroy pre-existing myelin and interfere with its formation.

In order to investigate the means by which ALA induces these effects, the authors checked the extent of oxidative damage to neural proteins and lipids. After twelve-days of ALA exposure, cells showed significantly increased levels of protein carbonyls and decreased levels of lipids (Felitsyn, 2009). As myelin formation is a highly energy-dependent process, they decided to check the expression of proteins involved in the electron transport chain. Interestingly, complexes III-V were unaffected while the expression of some proteins in Complexes I-II had decreased. In addition, over 50% of the ALA-treated ells contained mitochondria with altered polarity, indicating that the electron transport chain and ATP production are altered to some extent by this molecule. Based off these results, it appears that ALA is able to induce oxidative damage of both neuronal proteins and lipids. Whether this occurs by the oxidative properties of ALA or an increase in ROS production at the electron transport chain remains to be seen.

While the pathogenesis of hepatic VP symptoms remain a mystery, it has been known for a while that certain factors such as fasting, barbiturates, and alcohol can precipitate these effects. Recently, a link between blood glucose levels and heme biosynthesis via the proliferator-activator receptor γ coactivator 1α (PGC-1α) was established (Handschin, 2005).

Knowing that PGC-1α is involved in the switch from glucose metabolism to β-oxidation as the primary source of energy in the liver during starvation, Handschin and coworkers decided to investigate its role in the induction of porphyric attacks during fasting. First, the authors linked ALAS-1 transcription to PGC-1α by infecting murine liver cells with an adenovirus expressing this protein. Results indicate an increased expression of ALAS-1 mRNA after stimulation with PGC-1α (Handschin, 2005). Next, the authors established the impact of insulin and glucagon, the hormones responsible for regulating blood sugar, on the rate of ALAS-1 transcription in the hopes of finding a link to PGC-1α. As expected, glucagon was found to increase ALAS-1 mRNA levels (as this hormone would be high during times of fasting) while insulin was found to have the opposite effect (Handschin, 2005). In addition, PGC-1α knock-out mice did not experience the glucagon-induced increased in ALAS-1 expression, suggesting it is the link by which this hormone promotes heme metabolism.

Since the 1980s, patients experiencing porphyric attacks have been treated with IV infusions of glucose and insulin (Chen, 2009). In order to determine the means by which this works, the authors took a closer look at the role of insulin-mediated ALAS-1 deactivation. Handschin discovered that insulin deactivation of ALAS-1 transcription was linked to its ability to induce the phosphorylation of FOXO1. Since FOXO1 and PGC-1α form a complex capable of binding to the ALAS-1 promoter region, stimulating its transcription, phosphorylation of FOXO1, which results in its exportation from the nucleus, would impair this process (Handschin, 2005).

Interestingly, PGC-1α knock-out mice were also shown to be protected from models of porphyric attacks. First, the authors used the barbiturate phenobarbital and found that it was unable to induce elevated levels of ALAS-1 mRNA in these rodents. Next, they used lead and 1,4-dihydrocollidine to imitate porphyric attacks, as both substances are able to inhibit a variety of the enzymes involved. Results showed that ALAS-1 mRNA levels, plasma aminolevulinate, and plasma porphobilinogen levels did not match the elevated amounts characteristic for porphyric attacks (Figure 7).

Figure 7. Levels of ALAS-1 mRNA, ALA, and PBG under normal conditions and after stimulation with PbCl2 or DDC to imitate the porphyric condition. Source, Handschin, 2005.
Figure 7. Levels of ALAS-1 mRNA, ALA, and PBG under normal conditions and after stimulation with PbCl2 or DDC to imitate the porphyric condition. Source, Handschin, 2005.

In summary, the induction of porphyric attacks by fasting has been linked to the glucagon-mediated activation of ALAS-1 transcription by PGC-1α. In addition, the ability of insulin to resolve these symptoms has also been linked by its ability to inhibit PGC-1α activation of ALAS-1. While this provides insight into the means by which heme biosynthesis activation can induce VP symptoms, work that seeks to understand the means by which heme precursors induce these disease characteristics is necessary in order to better understand and treat this condition.