|Home | About | Journals | Submit | Contact Us | Français|
Heme is a cytotoxic, hydrophobic tetrapyrrole that crosses multiple biological membranes for incorporation into proteins critical for numerous biological processes. Thus, a prima facie argument can be made that heme trafficking within the aqueous cellular milieu must be mediated by specific intra- and intercellular pathways. Embryonic development in Caenorhabditis elegans, a heme auxotroph, is inextricably dependent on maternal heme acquisition. Here we show that HRG-3 is required to deliver maternal heme to oocytes for zygotic development. HRG-3 binds heme and is exclusively secreted by the intestine during heme insufficiency into the interstitial fluid for transport of maternal heme to extra-intestinal cells. HRG-3 deficiency results either in death during embryogenesis or in developmental arrest immediately post hatching – phenotypes that are fully suppressed by maternal but not zygotic hrg-3 expression. Our results establish an unprecedented role for HRG-3 as an intercellular heme chaperone in zygotic development and maternal-embryonic nutrition in C. elegans.
Heme-containing proteins are found in nearly all phyla of organisms (Hardison, 1996) and play essential roles in a wide range of biological process (Faller et al., 2007; Kaasik and Lee, 2004; Okano et al., 2010; Severance and Hamza, 2009). In mammalian cells, heme is either imported from the extracellular milieu through the plasma membrane (Uc et al., 2004; Worthington et al., 2001) or synthesized within the mitochondria for export to the cytoplasm for delivery to extra-mitochondrial compartments for insertion into a repertoire of hemoproteins (Dailey, 2002; Severance and Hamza, 2009). Free heme is an amphipathic planar macrocycle that can intercalate into membranes where it may promote damage to cellular macromolecules (Balla et al., 1991). Consequently, specific cellular pathways must exist for the directed transport, trafficking, and delivery of heme to numerous cellular destinations - but none have been found to date (Severance and Hamza, 2009). Previously, we identified the first bona fide metazoan heme importer HRG-1 (SLC48A1), which we propose plays a critical role in regulating cellular heme homeostasis in the roundworm Caenorhabditis elegans and vertebrates (Rajagopal et al., 2008). Heme export is mediated by a major facilitator superfamily protein, the feline leukemia virus subgroup C cellular receptor (FLVCR), in red blood cells and macrophages (Keel et al., 2008; Quigley et al., 2004). Hemopexin, a serum heme-binding protein, may facilitate heme export by physically interacting with FLVCR (Yang et al., 2010). Together, these proteins constitute part of a larger, intricate network to maintain organismal heme homeostasis – which heretofore remain poorly understood (Severance et al., 2010).
In an effort to identify additional components of the heme transport pathways, we took advantage of C. elegans, a heme auxotroph (Rao et al., 2005). In worms, nutritional heme is transported into the intestine by membrane bound permeases - HRG-1 and HRG-4 (Rajagopal et al., 2008). However, it’s unclear how tissues such as muscle, neurons, hypodermal cells, and embryos acquire heme from the intestine. Here we identify HRG-3, a novel heme-binding protein, which functions to transport heme from intestinal cells to extra-intestinal tissues including oocytes. Our results suggest that HRG-3 is an intercellular heme carrier that is essential for early development in C. elegans.
C. elegans wildtype N2 worms maintained axenically in mCeHR-2 liquid medium are gravid adults in 3.5 days in the presence of optimal concentrations of heme (20 µM) (Rao et al., 2005). However, their progeny are growth arrested at the fourth larval stage (L4) in the absence of supplemented heme. To differentiate the effects mediated by maternal heme from zygotic heme, we cultured parental worms (P0) at 1.5, 20, and 750 µM heme, all of which allow normal development and fertility, and the ensuing progeny (F1) were maintained at either 0 or 20 µM heme (Figures 1A and S1). Strikingly, when grown at 0 µM heme, F1 worms obtained from P0 mothers cultured at 1.5 µM heme were growth arrested at the first larval stage (L1), whereas F1 worms derived from P0 worms grown at 750 µM heme grew to young adults prior to becoming growth arrested. Irrespective of the P0 heme concentrations, F1 progeny developed normally when grown at 20 µM heme (Figure 1A and S1). These results suggest that larval development after hatching is dependent upon maternal (P0) deposition of heme, and that in the presence of heme, the F1 progeny can overcome maternally-induced heme deficiency.
To corroborate these results, we used a transgenic heme sensor strain in which the expression of intestinal GFP is inversely correlated with heme levels in the worm (Sinclair and Hamza, 2010). When worms were maintained at low concentrations of heme, strong GFP expression was observed both in the maternal intestine and in the embryos. However, embryonic GFP expression was severely attenuated, concomitant with maternal GFP, when mothers were provided with 20 µM heme, further demonstrating that heme levels in the embryos are linked to maternal heme status (Figures 1B and 1C).
A previous transcriptome analysis identified several hundred heme responsive genes (hrgs) in worms grown in axenic mCeHR-2 liquid culture supplemented with 4, 20, and 500 µM heme (Severance et al., 2010). To identify genes that may play a role in heme delivery, we first sorted genes based on the degree of upregulation under low heme conditions followed by three additional criteria. They should encode proteins (a) with a molecular weight of ≤30 kDa – a feature characteristic of metallochaperones (Kim et al., 2008), (b) with conserved amino acids which bind heme (H, Y, or C), and (c) which lack multi-span transmembrane domains. These criteria resulted in the identification of F58E6.7, which was upregulated >70 fold by low heme in the microarray. Northern blot analysis revealed the presence of a single ~370 nucleotide transcript (Figure 1D) and qRT-PCR revealed that worms grown in 1.5 or 4 µM heme increased the abundance of F58E6.7 mRNA by more than 900-fold and 400-fold, respectively, over what is found in worms grown in 20 µM heme (Figure 1E). Consistent with the Northern blotting results, 5’ and 3’ RACE confirmed the presence of a ~377 nucleotide mRNA containing a ~9 nucleotide 5’ untranslated region (UTR), three exons and a 155-nucleotide 3’ UTR (not shown). The ORF encodes a 70-amino acid protein with a predicted molecular mass of 8.1 kDa (Figure 1F). Within the amino terminus of F58E6.7 resides a stretch of hydrophobic amino acids, which could serve as either a transmembrane region or a signal peptide. BLAST searches and gene prediction algorithms identified putative homologs in other Caenorhabditis species (Figure 1F). These homologs share >50% sequence identity at the amino acid level. Consistent with genome nomenclature, we termed F58E6.7 as hrg-3.
To determine the function of HRG-3 in C. elegans, we analyzed worms containing a deletion in hrg-3. The tm2468 strain contains a 218-bp deletion that encompasses part of the promoter, the first two exons and the first intron, resulting in a null mutant (Figures 2A and 2B). These mutant worms have no overt phenotypes when fed the standard worm diet containing E. coli strain OP50 (not shown). Because hrg-3 is highly upregulated in worms grown at low heme conditions in mCeHR-2 liquid medium, and to rigorously analyze the hrg-3 mutant phenotype, we sought to recapitulate the heme deprivation conditions on agar plates with E. coli as the food source. Since the OP50 E. coli strain can synthesize heme endogenously, it was not possible to deplete the bacteria of heme. The bacterial strain RP523 is defective in hemB which encodes 5-aminolevulinic acid dehydratase (ALAD), the second enzyme in the heme biosynthesis pathway (Li et al., 1988). Consequently, RP523 is dependent upon exogenous heme for growth. By exposing worms to RP523 grown with different concentrations of heme, one can control heme levels in the worm via E. coli. Wildtype N2 worms exhibited a 1 to 2 day growth delay when fed RP523 grown with 1 µM heme compared to those grown on OP50, a growth phenotype that was not present when wildtype worms were fed RP523 grown with 10 to 50 µM heme (not shown). hrg-3 mutant worms, like wildtype worms, revealed the expected growth delay in the parental (P0) generation when fed RP523 grown with 1 µM heme. However, ~40% of eggs laid by the mutant worms failed to hatch (Figure 2C), and the F1 embryos that did hatch were growth arrested at the first larval stage (Figures 2D and 2E). The lethality and growth retardation phenotypes were completely rescued when hrg-3 mutants were fed RP523 that had been grown with 50 µM heme. Collectively, these results indicate that HRG-3 is essential for both embryonic and post-embryonic development under heme-limiting conditions and that the absence of HRG-3 results in heme deficiency that is manifested specifically in the F1 generation.
We determined the tissue and subcellular distribution of HRG-3. Worms expressing Phrg-3GFP transcriptional fusions had GFP in the worm intestine, with the greatest levels in the anterior (int2 and int3) and mid intestinal cells (int4–6). The anterior-most (int1) and the posterior-most gut cells (int7–9) possessed low levels of GFP (Figure 3A). GFP was only observed in worms that were maintained in ≤6 µM heme in mCeHR2 medium – consistent with the qRT-PCR results (not shown). Intestinal Phrg-3GFP expression was observed through all larval stages, and in both hermaphrodites and males (Figure S2). Zygotic expression of hrg-3 was first detected in late embryos at ~300 min of development (Figure S2).
To identify the subcellular distribution of HRG-3, we constructed transgenic worms which express the translational reporter Phrg-3HRG-3YFP. Worms grown in 2 µM heme possessed a weak HRG-3YFP signal that was located in cytoplasmic puncta within the worm intestine (Figure 3B, left panel). Unexpectedly, the majority of HRG-3YFP was present as vesicular structures outside the intestine in coelomocytes - macrophage-like scavenger cells located in the pseudocoelomic cavity (Figure 3B, right panel).
To determine whether the HRG-3YFP translational reporter was inadvertently expressed in extra-intestinal cells, we directed the expression of hrg-3 from the vha-6 promoter, a well-characterized intestinal promoter (Oka et al., 2001). Transgenic worms expressing Pvha-6HRG-3mCherry revealed strong HRG-3mCherry localization in extra-intestinal tissues including coelomocytes, the pseudocoelom, gonadal sheath cells, and the uterus (Figure 3C). Within intestinal cells, HRG-3mCherry signal was observed as distinct cytoplasmic vesicles that co-localized with mannosidaseGFP (Mans-GFP), a protein that localizes to the Golgi (Figure 3D). However, unlike HRG-3mCherry, expression of Mans-GFP from the vha-6 promoter showed no extra-intestinal localization (Figure 3D). Taken together, these results strongly suggest that HRG-3 is secreted from the intestinal cells into the pseudocoelom for uptake by extra-intestinal tissues.
To determine whether HRG-3 is a membrane anchored protein within an exosome or a soluble secreted protein, we synthesized truncated variants of HRG-3 that were tagged at the C-terminus with GFP. Expression of HRG-3-GFP in HEK293, a human kidney cell line, resulted in perinuclear localization. To examine the membrane orientation of HRG-3, we conducted fluorescence protease protection (FPP) assays (Lorenz et al., 2006). In this procedure transfected cells are sequentially exposed to digitonin to permeabilize the cells, followed by protease digestion to cleave cytoplasmic-located proteins. For example, the membrane protein prototype, CFP-CD3δ-YFP, which is targeted to the ER, contains a cytoplasmic YFP domain which is susceptible to protease digestion compared to the luminal CFP domain which remains intact (Figure 3E, upper panels). We found that HRG-3-GFP in transfected cells was not digested by the protease treatment, a result that was reproducible when the N-terminal 29 amino acids of HRG-3 (HRG-3N) was expressed as a YFP fusion protein (Figure 3E). These results indicate that the C-terminus of HRG-3 is protected from protease digestion and is not cytoplasmic.
To further identify the location of HRG-3 and the function of the N-terminal region, we synthesized truncated forms of HRG-3. Ectopic expression of these fusion proteins in HEK293 cells revealed that full-length HRG-3 and HRG-3N co-localized with the Golgi marker β 1,4- galactosyltransferase (GalT)-CFP, consistent with the localization of HRG-3 in the C. elegans intestinal cells (Figure 3F). However, deletion of the first 29 amino acids (HRG-3ΔN) resulted in a cytoplasmic localization, indicating that this N-terminal region is necessary for targeting HRG-3 to the secretory pathway. The localization of HRG-3 was not due to a large fluorescent protein tag because a HA epitope-tagged HRG-3 also colocalized with the GalT marker (Figure 3F, right panel), and co-expression of HRG-3-YFP and HRG-3-HA in the same cell resulted in both proteins co-localizing with the GalT marker (not shown).
To examine whether the N-terminus is cleaved or retained in HRG-3, transfected HEK293 cell lysates were subjected to analysis by SDS-PAGE and immunoblotting. The full-length HRG-3 and the HRG-3ΔN proteins were found to be equivalent in size (Figure 3G, lanes 1 and 2). Correspondingly, expression of HRG-3N-GFP resulted in a protein that was indistinguishable from GFP alone in size, suggesting that the N-terminal hydrophobic region is cleaved to produce the mature HRG-3-GFP protein (Figure 3G, lanes 3 and 4). To verify this result, we compared the molecular weight of HRG-3 that was generated either by in vitro transcription and translation or by transfecting HEK293 cells. Immunoblotting with anti-HA antibody revealed that the in vitro generated protein had a larger molecular weight, corresponding to the retention of the 25 amino acid leader peptide (Figure 3G, lanes 5 and 6). These observations suggest that the N-terminal portion of HRG-3 may be processed and removed in the mature protein.
To identify the membrane trafficking components which regulate HRG-3 secretion from the intestine, we used RNAi-mediated depletion of 45 genes which encode regulators of endocytosis and secretion (Table S1) (Balklava et al., 2007). We found that depletion of 15 trafficking factors caused HRG-3mCherry to either accumulate within the maternal intestine (vha-1), mislocalize in extra-intestinal tissues (sec-23) or embryos (sec-24.1) (Table 1 and Figure 4A). The majority of these candidate genes encoded for protein subunits that formed vesicle coatomer and vacuolar ATPase complexes (Table S1).
To determine whether HRG-3 secretion was tissue dependent, we ectopically expressed hrg-3mCherry in the hypodermis, specialized epithelial cells in C. elegans, using the dpy-7 promoter (Rolls et al., 2002). Transgenic worms expressing Pdpy-7HRG-3mCherry revealed HRG-3mCherry signal within cytoplasmic puncta in the hypodermis and in extra-hypodermal cells including coelomocytes, the pseudocoelom, and the uterus (Figure 4B and S3). As observed for the regulation of HRG-3 trafficking in the intestine, HRG-3 secretion from the hypodermis was also regulated by general membrane trafficking components (Figure 4A versus S3). Collectively, these results indicate that HRG-3 trafficking is mediated by general regulatory factors within the secretory pathway and is cell-type independent.
To examine whether HRG-3 secretion was dependent on organismal heme levels, we generated transgenic worms that expressed hrg-3YFP under the control of the inducible hsp-16.2 promoter, which is strongly expressed in the intestine and induced in response to heat shock. HRG-3YFP accumulated in the coelomocytes, which is indicative of secretion from the intestine, within 60 min after induction and continued to accumulate over time (Figures 4C and 4D). Importantly, Phsp-16.2HRG-3YFP transgenic worms accumulated similar amounts of HRG-3YFP irrespective of heme concentrations in the growth medium (Figure 4E). We were unable to examine HRG-3YFP secretion in worms grown at <1 µM heme because these animals were severely growth retarded.
To directly demonstrate that maternal HRG-3 is deposited within the embryo, we analyzed Pvha-6HRG-3mCherry mosaic transgenic worms in which the transgene was maintained as an extra-chromosomal array with a transmission efficiency of ~60%. Thus, P0 mothers that express HRG-3mCherry will lay F1 progeny that either express or lack the transgene (Figure 4F). Remarkably, 100% of F1 embryos isolated from transgenic mothers were positive for HRG-3mCherry even though 40% of these embryos did not express the transgene. HRG-3mCherry was visible at the time of gastrulation and detectable up to the mid-larval stages (L2 and early L3). Importantly, 100% of the F2 progeny, derived from non-transgenic HRG-3mCherry-positive F1 mothers, lacked any detectable HRG-3mCherry signal and the transgene (Figure 4F). These results confirm that maternal HRG-3 is transferred to all embryos irrespective of the zygotic genotype.
Our studies reveal that although hrg-3 is expressed in the intestine, hrg-3 loss-of-function mutants show embryonic lethality and growth retardation in the F1 generation when grown under heme insufficient conditions. These phenotypes could be due to HRG-3 deficiency either in P0 mother, in the F1 embryo, or both. To answer this question, we created a Phrg-3HRG-3ICSGFP construct in which hrg-3 and gfp were under the control of a single hrg-3 promoter but were separated by the SL2 intercistronic sequence (ICS) from rla-1 (Figure 5A). In C. elegans, the HRG-3ICSGFP transgene is transcribed as a single polycistronic mRNA, but yields two separate proteins: HRG-3 and GFP. Thus, GFP fluorescence is indicative of transgene expression (Figure 5A). Size and optical density analysis of stably transformed worms using a COPAS Biosort provided data that demonstrated that hrg-3 expression fully rescues the severe growth phenotype in the F1 progeny in hrg-3 deficient worms in the presence of low heme (Figure 5B).
To confirm that intestinal HRG-3 is crucial for heme delivery to extra-intestinal tissues, we used targeted gene rescue by expressing hrg-3 under the control of the intestine-specific vha-6 promoter. Unlike the hrg-3 promoter the vha-6 promoter is not heme regulated. Furthermore, to distinguish between maternal versus zygotic expressed HRG-3, we analyzed mosaic transformants in which the transgene transmission efficiency was ~40%. Thus, only P0 mothers that express HRG-3ICSGFP will lay progeny that either lack hrg-3 or express hrg-3 as an extra-chromosomal array. As expected, in the presence of low heme, >30% embryos from hrg-3 loss-of-function mothers failed to hatch and larvae that did hatch were growth arrested at the L1 stage. By contrast, <2% of embryos remained unhatched from P0 mothers expressing the hrg-3 transgene (Figure 5C). Importantly, a significant proportion of hatched embryos derived from hrg-3 expressing mothers continued to grow past the L2 stage, even though these larvae did not express hrg-3 (Figure 5D and Figure 5E, center and right panels). hrg-3 expressing embryos derived from crosses between hrg-3 loss-of-function mothers and HRG-3ICSGFP males were growth arrested. Only 2 out of 65 F1 progeny grew beyond the initial larval stages (Figure 5F and 5G). These data strongly suggest that targeted expression of hrg-3 from the maternal intestine is necessary and sufficient for embryonic development even when environmental heme is limiting.
Our results support a role for HRG-3 in heme delivery from the maternal intestine to oocytes. Based on this evidence, we postulated that when HRG-3 is available in limited quantities, greater heme accumulation would be found in the maternal intestine and a corresponding heme deficiency would exist in the developing embryos compared to wildtype worms. To estimate heme levels in these tissues, we crossed the heme sensor strain IQ6011 (Phrg-1GFP) with hrg-3 null mutants (Rajagopal et al., 2008; Severance et al., 2010; Sinclair and Hamza, 2010). The resulting IQ8011 gravid worms had reduced GFP levels in the intestine compared to IQ6011 worms (control) when grown in medium containing 1.5 or 2 µM heme (Figure 6A). Embryos derived from these mothers showed reproducibly higher levels of GFP, compared to wildtype controls (Figure 6B). Since the heme status is inversely correlated with the GFP expression in heme sensor worms, these results suggest that deletion of hrg-3 results in increased heme levels in the maternal intestine and reduced heme levels in the embryos. The modest differences in embryonic GFP levels between wildtype and hrg-3 embryos could be due to incomplete penetrance of the embryonic lethal phenotype (~40%; Figures 2C and and5C)5C) and environmental modifiers such as nutrient heme (Figure 2C). The consistently higher GFP (~5 fold) content in the intestine of the mother compared to the embryo could be attributed to the endoreduplication of chromosomes and multi-nucleation of the intestinal cells during worm development (Hedgecock and White, 1985). Taken together, our results show that HRG-3 deficiency causes perturbation of heme homeostasis in the maternal intestine and the embryo.
While the genetic and cell biology data are compelling in demonstrating that HRG-3 is involved in trafficking of heme from the maternal intestine to eggs, the data do not discriminate between direct or indirect functions of HRG-3 in heme homeostasis. To determine whether HRG-3 directly interacts with heme, we synthesized the mature secreted form of HRG-3 and measured its ability in vitro to bind heme. Pure HRG-3 is readily soluble in weak acidic solutions but becomes less soluble and gradually precipitates at neutral pH. However, addition of ferric protoheme to a solution of soluble HRG-3 at pH 7.0 resulted in a distinct spectroscopic peak at 416nm, while addition of ferrous protoheme resulted in a peak at 441nm (Figure 6C). These peaks are shifted and distinct from those of the free heme indicating that heme is definitively binding HRG3. Although we were unable to obtain an accurate association constant spectroscopically because the binding affinity of HRG-3 for heme was weak, titration of both ferric and ferrous heme revealed that, regardless of oxidation state, heme binds to HRG-3 at a stoichiometry of 1:2 (heme:protein) (Figure 6D). Notably, the soluble heme-bound HRG-3 slowly precipitates over several hours as a bright red complex indicating that the precipitated protein remains bound to heme with significant affinity.
As a heme auxotroph, embryonic and post-embryonic development in C. elegans is either dependent on maternal heme deposition (Figure 1A, upper panel), or larval heme acquisition (Figure 1A, lower panel). Our results uncover the crucial role of HRG-3 in maintaining embryonic heme homeostasis and its interdependence with maternal heme status (Figures 2C–E and Figures 5C–E). C. elegans acquires environmental heme through the coordinated functions of HRG-1 membrane-bound heme permeases located in the intestine (Rajagopal et al., 2008). Since a hermaphrodite worm has 959 somatic cells of which 20 are polarized intestinal cells (McGhee, 2007), the question remaining is how do extra-intestinal cells acquire heme? We postulate that this is partly accomplished through HRG-3, which we have shown above is likely to be an intercellular heme chaperone (Figure 6E). HRG-3 is transcriptionally upregulated in response to heme insufficiency, and secreted by the maternal intestine into the pseudocoelom, the worm’s circulatory system, for mobilization of heme to extra-intestinal tissues including the gonads and uterus. In the absence of HRG-3, heme accumulates in the intestine of gravid adults while the embryos are heme deficient resulting in embryonic lethality or growth arrest immediately after the embryos hatch.
When and how does HRG-3 transfer heme to the embryo? In C. elegans, oocyte fertilization results in a rapid assembly of a trilamellate chitinous eggshell by the time pseudocleavage of the one-cell embryo occurs (Johnston et al., 2006). The eggshell, which surrounds the developing embryo until hatching, provides a mechanical and osmotic barrier and ensures that early developmental events occur (Johnston et al., 2006). Given the impervious nature of the eggshell matrix to environmental factors, we speculate that heme deposition by HRG-3 must occur during oocyte maturation and prior to fertilization. This maternal-to-oocyte trafficking of heme exhibits striking similarity with the lipid transport pathways by vitellogenins, the major yolk precursor proteins. C. elegans contains six vitellogenins which are produced by the maternal intestine to bind lipids and translocated to the gonads via the pseudocoelom for yolk deposition in oocytes (Blumenthal et al., 1984; Kimble and Sharrock, 1983; Spieth and Blumenthal, 1985). However, unlike vitellogenins which are expressed only in the adult hermaphrodite (Blumenthal et al., 1984), hrg-3 is expressed during all developmental stages in both hermaphrodites and males (Figure S2). HRG-3 may, therefore, play a more extensive role than vitellogenins by mobilizing heme from intestinal cells to tissues other than embryos. Indeed, HRG-3 deficient F1 larvae are growth arrested in the presence of low heme implying that, in addition to in utero development which can be rescued by maternal HRG-3, sustained hrg-3 expression in the larvae is essential for post-embryonic development. A similar pathway may also exist for other metals such as zinc, which has been recently demonstrated to regulate meiotic maturation of mammalian oocytes and early embryonic development, implicating a role for zinc in the maternal legacy from egg to embryo (Kim et al., 2010).
What are the cellular factors which regulate HRG-3 trafficking? Of the 45 general regulators of membrane trafficking that were recently identified from a genome-wide RNAi screen for modulators of endocytosis and secretion of vitellogenin, RNAi depletion of 15 factors caused HRG-3 to accumulate or mislocalize in both the intestine and hypodermis (Table 1) (Balklava et al., 2007). Interestingly, these regulators broadly fall into two categories – coatomer complex and vacuolar ATPase subunits. HRG-3 trafficking and secretion may therefore be dependent on assembly of vesicles and its acidification. Although maternal HRG-3 persists from embryonic to larval stages, just like vitellogenin (Chotard et al., 2010), HRG-3 is not part of the vitellogenin complex because RNAi- depletion of all six vitellogenins did not alter HRG-3 secretion and trafficking (not shown). There are several examples of maternal contributions to the embryo that persist and function at later stages of development. For example, maternal cyclin E, a cell cycle checkpoint regulator, controls G1/S progression and coordinates cell proliferation and differentiation in C. elegans (Brodigan et al., 2003; Fay and Han, 2000). Maternal cyclin E is sufficient to regulate G1 cell cycle progression until the L3/L4 larval stages; cell cycle defects only become apparent when the maternal protein is exhausted in the F1 progeny.
Our biochemical studies demonstrate that the mature processed HRG-3 protein binds both ferrous and ferric heme with an apparent stoichiometry of two protein and one heme moiety. The spectroscopic data are consistent with a five coordinate high spin heme and reproducible by electron paramagnetic resonance spectroscopy (Albetel, Johnson, Dailey and Hamza, unpublished results). We propose that heme transfer by HRG-3 to target sites may be dependent on affinity gradients, as has been demonstrated for intracellular copper chaperones. Copper transfer is thermodynamically favored from low to intermediate to high affinity sites driven by intracellular metallochaperones; the hierarchy of copper binding amongst specific chaperones is governed by fast metal transfer, specific protein-protein recognition, and cellular compartmentalization (Banci et al., 2010).
Although intercellular transport of iron by the transferrin-transferrin receptor complex has been well documented, several lines of evidence also support the existence of an intercellular heme transport pathway in vertebrates. Firstly, even though knock-out of the heme synthesis pathway in mice is embryonic lethal, homozygous embryos survive at least till E3.5 suggesting the existence of heme stores (Magness et al., 2002; Okano et al., 2010). Secondly, zebrafish embryos with loss-of-function mutations in heme synthesis genes can survive from 10–25 days post fertilization (Childs et al., 2000; Dooley et al., 2008), plausibly because these embryos may either contain maternal-derived mRNA for heme synthesis enzymes or direct deposition of maternal heme. Thirdly, human patients with acute attacks of porphyrias, genetic diseases due to defects in heme synthesis enzymes, are administered heme intravenously as an effective therapeutic treatment which results in reducing heme synthesis intermediates and a concomitant increase in liver heme-dependent enzyme activities for cytochrome P450 and tryptophan 2,3-dioxygenase indicating that infused heme in the blood stream is utilized in toto by peripheral tissues (Puy et al., 2010) (Bonkovsky et al., 1991). Lastly, cell culture studies with human colon-derived Caco-2 cells and mouse macrophages reveal that a portion of heme, derived from dietary sources or senescent red blood cells, is released into the blood stream as an intact metalloporphyrin (Knutson et al., 2005; Uc et al., 2004). A potential candidate for an intercellular heme delivery protein would be hemopexin which scavenges heme with a Kd ~10−15 M and clears it from the circulatory system (Hrkal et al., 1974). The heme-hemopexin complex binds to the LRP/CD91 receptor, is endocytosed, and the majority of the hemopexin is degraded in the endo-lysosome of hepatocytes, macrophages, and syncytiotrophoblasts (Hvidberg et al., 2005; Tolosano et al., 2010). Surprisingly, hemopexin null mice are viable and fertile and present no evidence of tissue damage due to oxidative stress from abnormal heme and/or iron deposition under normal conditions; heme overload and hemolytic damage, however, causes tissue damage in these mice (Tolosano et al., 2010; Tolosano et al., 1999). We speculate that a functional homolog of HRG-3 may also exist in vertebrates as an alternate pathway to facilitate the targeted delivery and redistribution of heme between tissues and specific cell types and maintain systemic heme homeostasis.
Even though heme uptake and transport pathways are clearly conserved across metazoans (Severance and Hamza, 2009), heme auxotrophic organisms, such as C. elegans and parasitic helminthes, are crucially dependent on these pathways for utilization of environmental heme for growth and reproduction (Rao et al., 2005). Helminths affect more than a quarter of the world’s population (Chan et al., 1994; Hotez et al., 2008) and cause tens of billions of dollars of loss in animal and plant production annually (Fuller et al., 2008; Jasmer et al., 2003). Moreover, anthelminthics are becoming less effective in humans and livestock because of rampant drug-resistance (Fuller et al., 2008; Jasmer et al., 2003). We propose that an excellent anthelminthic target would be the HRG-3 mediated pathway for transporting heme to developing oocytes, especially in parasites such as hookworms, which infect more than a billion people worldwide and feed on host red blood cell hemoglobin (Held et al., 2006; Wu et al., 2009).
The heme-deficient E. coli strain RP523 was maintained in liquid LB medium supplemented with 1 µM heme at 37°C (Li et al., 1988). To prevent unequal growth of the RP523, overnight cultures were diluted into fresh medium with different concentrations of heme, grown for 5.5 h to OD600 of 0.2, and heat inactivated at 65°C for 2–5 min prior to seeding on 35 mm NGM agar plates. Synchronized L1 larvae of hrg-3 (tm2468) allele and its wildtype brood mates were placed onto RP523 plates and incubated at 20°C for 3–5 d. Five gravid hermaphrodites from each plate were allowed to lay eggs for 12–16 h on a new RP523 seeded plate. The embryos that did not hatch after 24–32 h were considered dead. The growth of F1 larvae was scored when the wildtype worms reached young adult stage. Those larvae that did not progress past L2 stage were considered growth arrested. DIC images were acquired on the F1 worms when the wildtype brood mates reached gravid stage.
Nematodes from plates containing RP523 were frozen when worms reached gravid stage. Worms from each sample (~100) were analyzed for length (time of flight) and optical density (extinction) using a COPAS BioSort (Union Biometrica, Holliston, MA). Gating parameters of time of flight 30–800 and extinction 15–800 were set by using synchronized L1s and mixed worm populations. Raw data outside this range were filtered to exclude particulates and bubbles. For rescue experiments, hrg-3 mutants expressing Phrg-3HRG-3ICSGFP (integrated) or Pvha-6HRG-3ICSGFP (extra-chromosomal array) were generated by genetic crosses. For zygotic rescue experiment, hrg-3 mutants were grown on RP523 with 1 µM heme for one generation, followed by crossing with male worms expressing Phrg-3HRG-3ICSGFP. The progeny were maintained at low heme for 5 days. Both worm size and GFP intensity were analyzed using COPAS BioSort for maternal and zygotic rescue experiments. The settings for GFP measurements in the zygotic rescue experiments were gain=3.5 and PMT voltage=750 but for all other experiments the settings were gain=3.0 and PMT voltage=600.
C. elegans transmitting the transgene Pvha6HRG-3mCherry with 60% efficiency to its progeny were grown on NGM plates seeded with RP523 supplemented with 4 µM heme. Embryos from adult gravid worms either with or without the transgene were released from the uterus using a needle. These embryos were analyzed for mCherry expression using a Leica DMIRE2 inverted microscope and Simple PCI software. In parallel, transgenic P0 gravid worms were individually picked onto new plates and allowed to produce F1 progeny and analyzed for mCherry expression by epifluorescence microscopy. These F1 worms were separated and grown till they lay F2 progeny which was analyzed for the mCherry transgene.
The procedure for fluorescence protease protection (FPP) assay was modified from the protocol by Lorenz et al. (Lorenz et al., 2006). HRG-3-GFP and control plasmid pCFP-CD3δ-YFP were transfected into HEK293 cells grown on Lab-Tek chambered coverglass (Nunc). After 24 h, the cells were washed with KHM buffer (110 mM potassium acetate, 2 mM MgCl2, and 20 mM HEPES, pH 7.3) and the cell chambers were moved to a DMIRE2 epifluorescence microscope (Leica) connected to a Retiga 1300 cooled Mono 12-bit camera. Time-lapse images were acquired before and after digitonin treatment (30 µM digitonin / 2 min), and following proteinase K (50 µg/ml) digestion.
All data are presented as mean ± SEM. Statistical significance was tested using one-way ANOVA followed by the Tukey-Kramer Multiple Comparisons Test in GraphPad INSTAT version 3.01 (GraphPad, San Diego). A P value of <0.05 was considered as statistically significant.
Additional material available in the Extended Experimental Procedures
We thank A. Golden and M. Krause for critical discussions and reading of the manuscript; B. Grant, D. Hall, and J. McGhee for insights into intestinal regulation in C. elegans; J. Lippincott-Schwartz for the FPP assay, B. Grant for RNAi clones and C. elegans strain RT1315; J. Hanover for use of the COPAS BioSort; T. Blumenthal for the SL2 ICS sequence; National Bioresource Project and S. Mitani for the hrg-3 strain. This work was supported by funding from the National Institutes of Health R01DK74797 (I.H.).
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.
Supplemental information. Supplemental Information includes Extended Experimental 2 Procedures, 2 tables, and 3 figures.
Author Contributions. C.C., T.K.S., J.S., and I.H. designed studies and interpreted data. H.A.D. conducted the biochemical heme binding studies. C.C. and I.H. wrote the manuscript. All authors discussed the results and commented on the manuscript.
Author Information The authors declare no competing financial interests.