Cuticle Integrity and Biogenic Amine Synthesis in Caenorhabditis elegans Require the Cofactor Tetrahydrobiopterin (BH4)

Abstract

Tetrahydrobiopterin (BH4) is the natural cofactor of several enzymes widely distributed among eukaryotes, including aromatic amino acid hydroxylases (AAAHs), nitric oxide synthases (NOSs), and alkylglycerol monooxygenase (AGMO). We show here that the nematode Caenorhabditis elegans, which has three AAAH genes and one AGMO gene, contains BH4 and has genes that function in BH4 synthesis and regeneration. Knockout mutants for putative BH4 synthetic enzyme genes lack the predicted enzymatic activities, synthesize no BH4, and have indistinguishable behavioral and neurotransmitter phenotypes, including serotonin and dopamine deficiency. The BH4 regeneration enzymes are not required for steady-state levels of biogenic amines, but become rate limiting in conditions of reduced BH4 synthesis. BH4-deficient mutants also have a fragile cuticle and are generally hypersensitive to exogenous agents, a phenotype that is not due to AAAH deficiency, but rather to dysfunction in the lipid metabolic enzyme AGMO, which is expressed in the epidermis. Loss of AGMO or BH4 synthesis also specifically alters the sensitivity of C. elegans to bacterial pathogens, revealing a cuticular function for AGMO-dependent lipid metabolism in host–pathogen interactions.

TETRAHYDROBIOPTERIN (BH4; 6R-5,6,7,8-tetrahydrobiopterin) is the natural cofactor of three distinct classes of enzymes including the aromatic amino acid hydroxylases (AAAHs), nitric oxide synthases (NOSs), and alkylglycerol monooxygenase (AGMO) (Werner et al. 2011). BH4 is therefore critical for a variety of cellular processes, being essential for the conversion of L-Phe to L-Tyr, for alkyl ether lipid metabolism, for synthesis of nitric oxide (NO), and synthesis of the neurotransmitters serotonin (5-hydroxytryptamine, 5HT) and dopamine (DA) and their derivatives. BH4 is present in many eukaryotes (Werner-Felmayer et al. 2002), including the nematode Caenorhabditis elegans (Calvo et al. 2008), in which the functions of the AAAHs phenylalanine hydroxylase (PAH, gene pah-1), tyrosine hydroxylase (TH, gene cat-2), and tryptophan hydroxylase (TPH, gene tph-1) are well established (Lints and Emmons 1999; Loer et al. 1999; Sze et al. 2000; Calvo et al. 2008). C. elegans lacks an endogenous NOS (Gusarov et al. 2013); as shown below, C. elegans encodes a single ortholog of the recently characterized AGMO (Watschinger et al. 2010).

In mammals, BH4 is synthesized de novo in four steps from GTP by at least three enzymes: GTP cyclohydrolase I (GTPCH1, human gene GCH1), 6-pyruvoyl tetrahydrobiopterin synthetase (PTPS, human gene PTS), and either sepiapterin reductase (SR), carbonyl reductase, and/or aldose reductase (Figure 1, Table 1; Werner et al. 2011). BH4 synthesis is regulated through the action of the GTPCH1 feedback regulatory protein (GFRP), known to mediate the activation or inhibition of mammalian GTPCH1 by L-Phe or BH4, respectively. In humans, mutations in the GCH1 gene can be recessive or cause a dominant Dopa-responsive dystonia, with or without hyperphenylalaninemia (HPA) (Ichinose et al. 1999). Mutations in the PTS gene lead to BH4-deficient HPA (also called atypical HPA or malignant phenylketonuria) (Thöny and Blau 1997). In BH4-deficient HPA patients, the neurological symptoms vary in severity depending on the degree of reduction in biogenic amine and nitric oxide levels. These conditions are manageable by carefully monitored biopterin supplementation and other treatments (Blau et al. 2001; Longo 2009). Pts knockout mice die within 48 hr if untreated with BH4 and neurotransmitter precursors (Sumi-Ichinose et al. 2001; Elzaouk et al. 2003).

Figure 1.

Biosynthesis, regeneration and utilization of tetrahydrobiopterin (BH4), and functions of BH4-dependent enzymes in C. elegans. Enzymes catalyzing pathway reactions are indicated by Roman numerals (Werner et al. 2011). (I) GTP cyclohydrolase I (E.C. 3.5.4.16); (II) 6-pyruvoyl tetrahydropterin synthase (E.C. 4.2.3.12); (III) sepiapterin reductase (E.C. 1.1.1.153), (IV), pterin-4a-carbinolamine dehydratase (E.C. 4.2.1.96); (V) [quinoid] dihydropteridine reductase (E.C. 1.6.99.7); (VI) phenylalanine hydroxylase (E.C. 1.14.16.1); (VII) tyrosine hydroxylase (E.C. 1.14.16.2); (VIII) tryptophan hydroxylase (E.C. 1.14.16.4); and (IX) alkylglycerol monooxygenase (E.C. 1.14.16.5). Names of C. elegans genes encoding these enzymes are shown in boxes adjacent to Roman numerals; gray boxes indicate genes for which knockout mutants are described for the first time in this work. Although cat-4 mutant phenotypes have been described previously, we demonstrate here that the gene encodes GTPCH1. Top left: Pathway for de novo BH4 synthesis. The gene encoding the enzyme catalyzing the final step(s) in BH4 synthesis is unknown. Bottom left: Pathway for BH4 regeneration. Right: Four enzymes that use BH4. Mutants in cat-4 and ptps-1 genes (BH4-deficient) have all phenotypes listed in the box at bottom. Mutants in individual BH4-dependent enzyme genes have the indicated subset of BH4-deficiency phenotype (dashed line).

Table 1. Pterin synthesis, regeneration, and related genes in C. elegans.

Protein	C. elegans gene name	Genetic location / Genomic position (coding)	BLAST2 Expected-value vs. human and fly proteins*	Mutant allele(s)
GTP cyclohydrolase I (GTPCH1)	cat-4 / F32G8.6	V: 2.59 V: 10,564,851-10,567,502 bp	Hs: 7 × 10−84 Dm: 6 × 10−81	e1141, e3015, e3030, ok342, tm773
Pyruvoyl tetrahydropteridine synthase (PTPS)	ptps-1 / B0041.6	I: -1.03 I: 4,652,907-4,652,187 bp	Hs: 9 × 10−39 Dm: 8 × 10−44	e3042, tm1984
Sepiapterin reductase (SR)	No ortholog	NA	**Best matches in Ce by BLASTP Hs SR: 6 × 10−9 Dm SR: 9 × 10−12	NA
Carbonyl reductase (CR)	No ortholog	NA	**Best match in Ce by BLASTP Hs CBR1: 2 × 10−12	NA
Aldose reductase (AR)	Y39G8B.1	II: 21.67 II: 13,970,583-13,972,946 bp	Hs AKR1B1: 2 × 10−97 Hs AKR1C3: 1 × 10−87	ok1682
Pterin carbinolamine dehydratase (PCBD)	pcbd-1 / T10B11.1	I: 1.57 I: 6,951,134-6,951,946 bp	Hs: 9 × 10−39 Dm: 2 × 10−37	tm5924
Quinoid dihydropteridine reductase (QDPR)	qdpr-1 / T03F6.1	III: 21.21 III: 13,393,783-13,394,837 bp	Hs: 3 × 10−70 Dm: 3 × 10−77	tm2337, tm2373
Dihydrofolate reductase (DHFR)	dhfr-1 / C36B1.7	I: 3.04 I: 8,736,060-8,737,028 bp	Hs: 3 × 10−34 Dm: 3 × 10−36	None
GTP cyclohydrolase I feedback regulatory protein (GFRP)	gfrp-1 / Y38C1AA.13	IV: -26.81	Hs: 5 × 10−28 Dm: no GFRP	None
		IV: 203,882-207,984 bp

*BLAST2: For multiple isoforms, the score shown is the best match between a Ce predicted protein and Hs (human) or Dm (fruit fly) protein. **Best match of sepiapterin reductase (SR) or carbonyl reductase (CR) via BLASTP to Ce proteins (nr database). Both CR and aldose reductase are possible partial substitutes for SR. NA, not applicable.

As well as being required for neurotransmitter synthesis, BH4 and its derivatives are important for the synthesis of pigments and quinones involved in cross-linking external cuticle layers in invertebrates (Iino et al. 2000; Kato et al. 2006). The molecular genetics of biopterin synthesis and biogenic amine metabolism have been extensively characterized in Drosophila melanogaster (Wright 1987; O’Donnell et al. 1989). In Drosophila, where dopa and dopamine are required for sclerotization and tanning of the cuticle (Neckameyer and White 1993), knockout Punch (GTPCH1) and purple (PTPS) mutants die as embryos due to severe cuticle abnormalities and/or a requirement for serotonin in germband extension (Mackay and O’Donnell 1983; Colas et al. 1999). Interestingly, GFRP is not found in Drosophila (Funderburk et al. 2006).

When BH4 is used by the AAAHs and AGMO in their respective hydroxylation reactions, it is oxidized to pterin 4-a-carbinolamine (Figure 1). This oxidized biopterin can be regenerated in mammals to BH4 by two reduction steps catalyzed by pterin carbinolamine dehydratase (PCBD) and quinoid dihydropteridine reductase (QDPR) (Werner et al. 2011). In humans, mutations in QDPR have severe effects like those in GCH1 and PTS genes, whereas mutation in the human PCDB1 gene yields a milder clinical picture (Opladen et al. 2012). BH4 can also be oxidized nonenzymatically and subsequently regenerated to BH4 by dihydrofolate reductase (DHFR) (Werner et al. 2011; Xu et al. 2014).

Analysis of the complete genomic sequence of C. elegans reveals orthologs of genes encoding biopterin synthesis, regulation, and regeneration enzymes known from other animals (Figure 1 and Table 1; see also Hobert 2013). C. elegans has clear orthologs encoding the first two BH4 synthetic enzymes, GTP cyclohydrolase I (GTPCH1, gene cat-4) and 6-pyruvoyl-tetrahydropterin synthase (PTPS, gene ptps-1). C. elegans appears to lack an ortholog of sepiapterin reductase (Kallberg et al. 2002), but does encode an aldose reductase (Table 1), which can partially substitute for SR in mammals (Park et al. 1991; Iino et al. 2003). The C. elegans genome also contains genes encoding biopterin regeneration enzymes PCBD, QDPR, and DHFR (genes pcbd-1, qdpr-1, and dhfr-1) and a clear ortholog of GFRP (gfrp-1).

Although cat-4 mutants were isolated in 1975 (Sulston et al. 1975), the biochemical genetics of BH4 in C. elegans has not been previously examined. cat-4 mutants were found based on their lack of the catecholamine DA (Sulston et al. 1975) and were subsequently found to be 5HT deficient (Desai et al. 1988). These neurotransmitter deficiencies in cat-4 mutants cause a variety of subtle behavioral abnormalities, including defective locomotory rate regulation and male mating (Loer and Kenyon 1993; Sawin et al. 2000). We and others have also found that cat-4 mutants are hypersensitive to a variety of agents, suggesting they might have a generally “leaky” cuticle (Loer 1995; Weinshenker et al. 1995; Cronin et al. 2005; Baker et al. 2012). A mechanistic explanation for the cuticle defects in cat-4 mutants, however, has been lacking.

Here we characterize the C. elegans pathway for BH4 biosynthesis, regulation, and regeneration. We find that cat-4 and ptps-1 mutants are biopterin-, 5HT- and DA-deficient and lack GTPCH1 and PTPS activities, as predicted. BH4-deficient animals have a fragile cuticle that is more permeable to small molecules, resulting in hypersensitivity to multiple chemicals. As deletion mutants in the AAAH genes do not display cuticle fragility or chemical hypersensitivity, we inferred that these phenotypes might reflect impaired function in another BH4-dependent enzyme. We show here that loss of function in the biopterin-dependent lipid metabolic enzyme AGMO (Watschinger et al. 2010) results in chemical hypersensitivity and cuticle fragility like that observed in BH4-deficient mutants. We find that agmo-1 is expressed in the epidermis, consistent with its requirement for a BH4 cofactor and its role in cuticle integrity. Furthermore, agmo-1 and the BH4-deficient mutants share a common phenotype of sensitivity to bacterial infection by Leucobacter Verde1. Our studies provide the first in vivo evidence for a role for AGMO in epidermal lipid metabolism and in pathogen defense, with implications for the function of this enzyme in other animals and in humans.

Materials and Methods

C. elegans culture, strains, and transgenes

Routine culturing of C. elegans was performed as described (Brenner 1974); strains were grown at 20° for all experiments, although acute analyses were at room temperature (RT) (21°–23°). Nomenclature used for C. elegans genetics conforms to conventions (Horvitz et al. 1979). Worm strains used are listed in Supporting Information. Deletion mutant strains of cat-4, ptps-1, qdpr-1, pcbd-1, and agmo-1 were outcrossed three to five times before most analyses. Homozygosity of deletion strains was confirmed by PCR.

We generated transgenes for cat-4, ptps-1, pcbd-1, qdpr-1, gfrp-1, pah-1, and agmo-1 by the duplex PCR method (Hobert 2002); see Table S1 on transgenes and transgenics. Amplified genomic DNAs were fused to amplified GFP; duplex products were co-injected with Pttx-3::RFP marker plasmid into wild type to generate transgenics. We also examined reporter transgenics previously described by others (Table S2).

The cat-4 mutant allele sequencing and bioinformatics

The cat-4(e1141) mutation was identified by PCR-amplifying exons from genomic CB1141 DNA, then Sanger sequencing the purified PCR product as described (Hare and Loer 2004). The mutation was confirmed by sequencing both strands for two independent PCR reactions. The cat-4(gk245686) mutation (strain VC20144) was confirmed in the same manner. PCR and sequencing primers were designed using Primer3 (http://www-genome.wi.mit.edu/cgi-bin/primer/primer3.cgi).

C. elegans genomic and predicted cDNA sequences were retrieved from WormBase and/or GenBank. Blast searches and Blast2 comparisons were performed using the NCBI Blast server. Multiple sequence alignments used CLUSTALW (www.ebi.ac.uk/clustalw/).

Detection of serotonin and dopamine in situ

Serotonin was assessed in whole-mount worms by immunofluorescence as previously described (Desai et al. 1988; Loer and Kenyon 1993) using a polyclonal antiserum against 5HT paraformaldehyde conjugated to BSA (Sigma, St. Louis; catalog no. E5545, lot 091K4831) previously tested for specificity (Loer and Rivard 2007). DA was assessed in whole-mount worms via formaldehyde-induced fluorescence (FIF) as previously described (Lints and Emmons 1999). Treatment of worms with 5-hydroxytryptophan (5-HTP) or L-dopa prior to staining was done as described (Rivard et al. 2010).

Chemical sensitivity and cuticle integrity tests and male mating assays

Most chemicals were obtained from Sigma-Aldrich, including levamisole (T512), 5-hydroxytryptophan, and L-dopa. BH4 and sepiapterin were obtained from B. Schircks Laboratories (Jona, Switzerland). We tested adult hermaphrodites grown for 3 days at 20° from synchronized L1 worms hatched in M9. Eggs isolated by bleaching gravid adult hermaphrodites were washed extensively with M9, then hatched in M9 at RT (22°–23°). Starved L1 worms were used immediately or stored at 12°–13°. Synchronized L1s were used no later than 7 days after egg isolation. Adult worms were washed from plates with M9 and 500-µl aliquots placed in 24-well plates. Because of “swimming-induced paralysis,” which can vary based on biogenic amine levels (McDonald et al. 2007), wells were scored after approximately the same length of time in M9 buffer. Chemicals (SDS, levamisole) were made as 2× solutions in M9; 500 µl of 2× solution was added to a well with 500 µl worms (50–100) to yield a working concentration. Individual worms in a well were observed through a stereomicroscope and scored as immobile if no movement was seen during the brief moment of observation (≤1 sec).

Rapid bleach hypersensitivity tests were done with standard C. elegans “alkaline bleach” [4.5% sodium hypochlorite (household bleach)/1M NaOH]. Gravid adult hermaphrodites were tested on NGM plates by applying a ∼5-µl drop of bleach on a worm and assaying time to rigidity using a stereomicroscope. Cuticle disintegration assays were carried out as described (Calvo et al. 2008). Briefly, gravid adults placed in a milder alkaline bleach solution (1% sodium hypochlorite, 0.25 M NaOH) were observed under a dissecting stereomicroscope, and the time of first major break in the cuticle was recorded. Plates were agitated manually every 30 sec. A total of 15–30 worms were scored in every experiment.

Preparation of worm homogenates and the effect of BH4 supplementation

Nematode cultures were grown on NGM plates supplemented with 5 mM ascorbate, 200 μM BH4 for 3–5 days (until food was almost depleted). Mixed-stage populations (nonsynchronized, with predominance of adult worms) were recovered from plates with sterile M9 buffer and cleaned of bacteria by successive centrifugations at 1000 rpm. Finally, worms were incubated 30 min in M9 with agitation to reduce bacteria in the gut and centrifuged again; the resulting pellet was immediately frozen in liquid nitrogen for later use. The frozen worm pellet was resuspended in 400 µl distilled water containing 5 mM dithioerythritol, homogenized with an Ultraturrax (Iba, Stauffen, Germany), frozen in liquid nitrogen, thawed, homogenized again, and centrifuged 5 min at 16,000 × g at 4°.

Determination of GTPCH1 and PTPS activity and BH4 content

GTPCH1 and PTPS enzymatic assays were carried out as described (Werner et al. 1997; Heller et al. 2001), with modifications to reduce the amount of required material. Results were related to protein concentrations measured in homogenates and eluates by Bradford assay (Biorad, Vienna). For BH4 determinations, 50 µl of homogenate was used for both iodine oxidation in acid and alkaline medium assays (Heller et al. 2001), and BH4 concentration was calculated from the difference of resulting biopterin from the two oxidation schemes (Fukushima and Nixon 1980). For acidic oxidation, 50 µl homogenate was mixed with 5 µl 1 M HCl and 5 µl 0.1 M iodine solution (prepared in 0.25 M potassium iodide). For basic oxidation, HCl was replaced by 1 M NaOH. After 60 min at RT in the dark, 10 µl 1 M HCl was added to the alkaline oxidation only, both incubations were centrifuged for 5 min at 16,000 × g at 4°, and supernatants added to 10 µl freshly prepared 0.1 M ascorbic acid. Biopterin was measured by HPLC after injection of 10 µl on a Nucleosil 10 SA column (250 mm long, 4 mm inner diameter, 10 µm particle size; Macherey Nagl, Düren, Germany), elution with 50 mM potassium phosphate, pH 3.0 at 35° and fluorescence detection (Excitation 350 nm, Emission 440 nm) with an Agilent 1200 HPLC (Agilent, Vienna).

For determining GTPCH1 activity, 80 µl of homogenate was separated from low MW compounds using Micro Bio-Spin 6 columns (Biorad) equilibrated to GTPCH1 assay buffer (100 mM Tris-HCl, pH 7.8, 2.5 mM EDTA, 300 mM KCl, 10% (v/v) glycerol), incubated with 1.5 mM GTP for 90 min at 37° in 85.7 µl total volume. Reaction was stopped by addition of 2.85 µl 1 M HCl and 2.85 µl 0.1 M iodine. Oxidation of resulting 7,8-dihydroneopterin triphosphate to neopterin triphosphate was achieved by incubation for 60 min at RT in the dark. After 2 min centrifugation at 16,000 × g at 4°, 2.85 µl 0.1 M ascorbic acid was added to the supernatant. After neutralization by addition of 2.85 µl 1 M NaOH, neopterin phosphates were cleaved to neopterin by incubation with 6.4 units alkaline phosphatase for 30 min at 37°. Neopterin was quantified by HPLC using 10 µl final reaction mixture injected on a reversed phase C-18 column (250 mm long, 4 mm inner diameter, 5 µm particle size; Lichrosphere, Merck, Darmstadt, Germany), eluted with 15 mM potassium phosphate, pH 6.0, at 25° and fluorescence detection (Ex 350 nm, Em 440 nm).

For measuring PTPS activity, 80 µl of homogenate was separated from low MW compounds using Micro Bio-Spin 6 columns (Biorad) equilibrated to PTPS assay buffer (0.1 M Tris-HCl, pH 7.4, 20 mM MgCl₂), and incubated with 40 µM freshly prepared 7,8-dihydroneopterin triphosphate (using recombinant Escherichia coli GTPCH1) and 2 mM NADPH in the presence of E. coli-expressed recombinant mouse SR (2 nmol/min) in 100 µl total volume for 1 h at 37° (Werner et al. 1997). Reaction was stopped by addition of 5 µl 1 M HCl, and 5 µl 0.1 M iodine. Following further incubation for 1 h at RT in the dark, resulting biopterin was quantified by HPLC as for neopterin in the GTPCH1 assay described above.

Isolation of Leucobacter Verde1 resistance mutants and whole genome sequencing

Selections for increased resistance to Leucobacter Verde1 exploited the hypersensitivity of certain bus and srf mutants to this bacterial pathogen, which is completely lethal to such mutants but not to wild-type C. elegans (Hodgkin et al. 2013). Mutants in three such genes (bus-10, srf-2, and srf-5) were used initially. Populations of these mutants were mutagenized with 0.05 M EMS (Brenner 1974). After mutagenesis, 50–80 individual L4 worms were picked to separate plates and grown on E. coli OP50 for two generations. About 200 F₂ progeny from each plate were then transferred to mixed E. coli/Verde1 (10:1) lawns and incubated at RT for a further 7–10 days, after which most (273/300) plates contained only dead or dying worms. Plates with surviving fertile worms were retained, and single worms were picked from these to establish independent resistant lines. Outcrossing established that most resistant lines carried recessive extragenic mutations conferring resistance to Verde1 and increased bleach sensitivity. Genetic mapping and complementation tests, utilizing the bleach sensitivity phenotype, assigned 27 independent mutations to nine complementation groups. The largest complementation group (eight alleles) was mapped genetically to the right arm of LGIII. DNA from a mutant strain carrying one of these mutations (e3016) was prepared for whole genome sequencing using a standard library preparation without amplification, followed by Illumina 50-bp paired end sequencing on the HiSeq2000. Candidate mutations were identified using MAQGene (Bigelow et al. 2009), which revealed a nonsense mutation (W130opal) in agmo-1. Specific sequencing of agmo-1 in strains for other alleles in this complementation group showed that all carried predicted severe mutations in this gene. Similarly, two other complementation groups were found to correspond to cat-4 (alleles e3015 and e3030) and ptps-1(e3042). Detailed analyses of six other complementation groups will be presented elsewhere.

Results

The C. elegans cat-4 (catecholamine defective) gene encodes GTP cyclohydrolase I (GTPCH1)

The cat-4 gene maps genetically near a predicted GTPCH1-encoding gene (F32G8.6) on chromosome V. Sequencing F32G8.6 exons from the original allele cat-4(e1141), revealed a missense mutation that changes an amino acid (T66I) that is 100% conserved in GTPCH1 proteins (Figure 2, Figure S1). Two putative knockout mutants with deletions (ok342 and tm773) in F32G8.6 failed to complement e1141, were also 5HT- and DA-deficient, and were hypersensitive with a fragile cuticle (Table 2). We examined additional cat-4 mutants with missense mutations altering different highly conserved amino acids (Figure 2); these mutants showed a range of phenotypes. cat-4(e3030) mutants had little or no apparent 5HT or DA, whereas e3015 mutants had reduced 5HT and DA.

Figure 2.

The cat-4 gene encodes GTP cyclohydrolase I (GTPCH1). (A) Maps of cat-4 region with mutant alleles and gene structure. cat-4(e1141) maps genetically between sma-1 and unc-23 on chromosome V in the same region as gene F32G8.6. Nature of cat-4 alleles is shown below and approximate locations are indicated with arrows (point mutations) or red bars (deletions). Sequencing of cat-4 cDNAs confirms the gene model shown (Figure S1) and the predicted proteins used for alignments. Image is partly derived from the WormBase genome browser. (B) Alignment of C. elegans CAT-4 with GTPCH1 proteins from other metazoans. Asterisks below alignment show 100% conserved residues; colon indicates conserved highly similar residues; and period indicates conserved weakly similar residues. Green triangles indicate amino acids likely within the active site (Maita et al. 2004). Residues altered in cat-4 mutant alleles are marked with red letters/yellow backgrounds. Species abbreviations are as follows: Cel, C. elegans; Dme, Drosophila melanogaster; Spu, Strongylocentrotus purpuratus; Cin, Ciona intestinalis; Dre, Danio rerio; Xla, Xenopus laevis; has, Homo sapiens; Nve, Nematostella vectensis (a partial sequence); and Tad, Trichoplax adherens.

Table 2. Phenotypes of C. elegans pterin-related gene mutants.

C. elegans strain	Alleles tested	Serotonina	Dopamineb	Bleach hypersensitivity
N2 (wild type)	NA	+c	+	Non Hyp
cat-4(−) [GTPCH1]	e1141, ok342, tm773, e3030	–d	–	Hyp
ptps-1(−)	tm1984, e3042	–	–	Hyp
pcbd-1(−)	tm5924	+/–	+/–	Non Hyp
qdpr-1(−)	tm2337, tm2373	+/–	+/–	Non Hyp
tph-1(−)	e	–	+	Non Hyp
cat-2(−) [TH]	e	+	–	Non Hyp
bas-1(−) [AADC]	e	–	–	Non Hyp
pah-1(−)	e	+	+	Non Hyp
agmo-1(−)	e3016, e3019, e3029, e3047	+	+	Hyp
Y39G8B.1(−) [AR]f	ok1682	+	+	Non Hyp

NA, not applicable.

^aTested by immunoreactivity (IR).

^bTested by formaldehyde induced fluorescence (FIF).

^cWild type.

^dDeficient.

^eReported previously.

^fY39G8B.1 encodes an aldose reductase (AR) ortholog, a possible partial substitute for sepiapterin reductase. The putative mutant, however, shows no phenotypes consistent with BH4 deficiency.

Bleach hypersensitivity and reduced melanin phenotypes of cat-4(tm773) mutant worms are rescued in transgenics carrying a plasmid containing F32G8.6 coding sequence, 3′-UTR, and 1500 bp upstream (Baker et al. 2012). We tested these transgenic worms for 5HT and DA expression; in two independent transgenic lines carrying extrachromosomal arrays, a high percentage of worms were rescued for DA and 5HT in all normal dopaminergic and serotonergic neurons (Figure S2).

Mutants defective in BH4 synthesis display biogenic amine deficiency, chemical hypersensitivity, and cuticle fragility

We confirmed the role of a predicted PTPS gene (B0041.6; ptps-1; Figure S3) by examining the phenotypes of worms with a deletion (tm1984) or nonsense mutation (e3042) in coding sequence. We found that both ptps-1 mutants were 5HT-deficient, DA-deficient, and hypersensitive with a fragile cuticle, just like cat-4 mutants (Figure 3, Figure 4, Table 2). Also like cat-4 mutants, ptps-1 mutants can be rescued for 5HT immunoreactivity (5HT-IR) by treating worms with 5-HTP, the immediate precursor to 5HT and product of TPH activity, and rescued for DA (as seen by FIF) by treatment with L-dopa, the immediate precursor to DA and product of TH activity (Figure 3, C, F, I, and L). Both treatments bypass the need for TH and TPH function in synthesis of the neurotransmitters, and also show that serotonergic and dopaminergic neurons are present and appear morphologically normal in cat-4 and ptps-1 mutants. We also determined that ptps-1 mutant male worms, like cat-4 mutant males (Loer and Kenyon 1993), were defective in the “turning” step of male mating behavior (Figure S4).

Figure 3.

Neurotransmitter phenotypes of cat-4, ptps-1, and qdpr-1 mutants. (A–F) Anti-5HT immunofluorescence of adult worm heads. One prominent 5HT neuron cell body (neurosecretory motorneuron, NSM) is marked with an arrow. Anterior is to the left. (A) Wild type (N2) and (D) qdpr-1 mutants have normal 5HT; (B) cat-4 and (E) ptps-1 mutants lack 5HT. (C and F) 5HT-IR is restored in cat-4 and ptps-1 mutants by treatment with 5-hydroxytryptophan (5-HTP). 5-HTP also causes dopaminergic neurons to make 5HT. [Rescue of cat-4 mutants by 5-HTP has been shown previously (Loer and Kenyon 1993).] (G–L) FIF showing dopamine (DA) in larval worm heads. DA-containing cells have characteristic blue-green fluorescence; background is more yellow-green. DA cells in the head are indicated with arrows (CEPD and CEPV, smaller arrowhead; and ADE, broad arrowhead). Asterisks indicate nonspecific intestinal fluorescence. In some heads, one can see both right and left side DA neurons. Anterior is to the left. (G) Wild type (N2) and (J) qdpr-1 mutants have normal DA; (H) cat-4 and (K) ptps-1 mutants lack DA. (I and L) DA fluorescence is restored in cat-4 and ptps-1 mutants by treatment with L-dopa.

Figure 4.

cat-4, ptps-1, and agmo-1 mutants share a common hypersensitivity to exogenous chemicals. (A) Hypersensitivity and cuticle fragility shown by rapid death and disintegration of cat-4, ptps-1, and agmo-1 mutants in standard alkaline bleach. Adult gravid hermaphrodites were photographed after addition of bleach drop. Top panels: N2 (wild type) worms were still wriggling after 10 sec in bleach, whereas cat-4(tm773), ptps-1(tm1984), and agmo-1(e3016) mutants were immobile/dead in a few seconds (<5 sec). Middle panels: N2 worm is dead but intact 30 s after bleaching. Mutants have exploded at multiple sites by 15–30 sec, releasing internal contents. Bottom panels: N2 worm has ruptured, but remains largely intact at 2 min; mutant worms have completely disintegrated and cuticles vanished, leaving eggs and internal debris. (B and C) Cuticle fragility of cat-4, ptps-1 (B), and agmo-1 (C) mutants. Time (mean ± SD) for cuticle disintegration in mild alkaline bleach, scored for the first major break in the worm cuticle. Groups for each experiment compared with one-factor ANOVA followed by planned pairwise comparisons made with Scheffè’s F-test; each experiment showed significant differences among the groups (overall ANOVA, P < < 0.0001). (B) Representative experiment with worms from mixed stage cultures (n = 15); cat-4 and ptps-1 compared to wild type (N2). Asterisks (*) indicate significant differences (F-test, P < < 0.0001) between each mutant and wild type. Mutants were not significantly different from one another (P > 0.05). (C) Two agmo-1 mutants compared to wild type (N2), synchronized gravid adult hermaphrodites (n = 17–27). Both mutants were significantly different from wild type (P < <0.0001); double asterisks (**) indicate e3016 was also significantly different from e3047 (P = 0.002).

The hypersensitivity and cuticle fragility phenotypes of cat-4 and ptps-1 mutants are seen most dramatically when worms are placed in a standard “alkaline bleach” solution typically used to kill and dissolve gravid adult worms to isolate the more bleach-resistant eggs. Whereas wild-type worms die and become rigid after many seconds, and take minutes to dissolve, we observed that cat-4 and ptps-1 mutant adult worms died in a few seconds or less, ruptured quickly, and their cuticles dissolved completely in <2 min (Figure 4A). To quantify the cuticle fragility phenotype of cat-4 and ptps-1 mutants, we tested worms using a mild alkaline bleach treatment. Similar to the effect of standard bleach treatment, cat-4 and ptps-1 worms ruptured more quickly: a major break in the cuticle occurred after about 4 min in the solution, compared with 8 min for wild-type worms (Figure 4B). cat-4 and ptps-1 mutants showed comparably increased sensitivity to various chemicals and drugs, including the detergent SDS and the anthelminthic levamisole, which in nematodes acts as a cholinergic agonist (Figure S5, A and C). In tests of acute exposure causing worm immobility, the biopterin synthesis mutants were approximately twice as sensitive to SDS compared to wild type, and 5–20 times more sensitive to levamisole.

To demonstrate that CAT-4 and PTPS-1 are bona fide BH4 synthetic enzymes, we measured GTPCH1 and PTPS activity in soluble protein extracts from homogenized worms, as well as total biopterin and BH4 content. cat-4 null mutants lacked GTPCH1 activity, but displayed PTPS activity, whereas ptps-1 mutants lacked PTPS activity, but had normal GTPCH1, confirming loss of the predicted functions of these genes (Figure 5, A and B). In addition, very low levels of total biopterin and BH4 were detected in the mutants, compared with levels in wild-type worms (Figure 5, C and D). As expected, BH4 supplementation did not alter enzymatic activities of GTPCH1 or PTPS (Figure 5, A and B), which were determined in homogenates freed from low molecular weight compounds by gel filtration. Rearing worms with exogenous BH4 also failed to increase levels of BH4 in mutants, although total biopterins did increase, indicating some uptake of biopterin (Figure 5, C and D). We observed no rescue of cuticle fragility (Figure S8) or neurotransmitter synthesis in cat-4 and ptps-1 mutants supplemented with exogenous biopterins.

Figure 5.

Biopterin synthetic enzyme activity and biopterin levels in wild-type and mutant C. elegans. Enzyme activity or biopterin content from worm homogenates from mixed stage cultures. Untreated (-/-), with ascorbate (asc, 5 mM) alone, or asc plus BH4 (200 µM), as indicated below each column. Enzyme activities and BH4 levels from worm homogenates were determined by HPLC with fluorescence detection (see Materials and Methods). All levels are expressed relative to protein mass. Mean ± SEM are shown for three to four independent measurements (two to three for cat-4 mutant). (A) GTPCH1 activity. (B) PTPS activity. (C) Total biopterin derivatives concentration. (D) BH4 concentration. GTPCH1 and PTPS activities were measured in soluble fraction of worm extracts. Total biopterin and BH4 measurements were taken from the same samples. Worm strains: N2, wild type; cat-4(tm773); and ptps-1(tm1984).

The chemical hypersensitivity and cuticle fragility defects of biopterin mutants result from loss of function in alkylglycerol monooxygenase AGMO

As shown previously, knockout mutants in individual AAAH genes (pah-1, cat-2, and tph-1) are normal with respect to chemical sensitivity and cuticle strength (Table 2); pah-1 mutants display cuticle defects only in sensitized backgrounds in which the cuticle is already compromised (Calvo et al. 2008). These observations suggested that deficiency in function of the newly identified BH4-dependent enzyme AGMO might explain the hypersensitivity and cuticle fragility phenotypes of cat-4 and ptps-1 mutants. We therefore examined other mutants with similar phenotypes of cuticle fragility. Genetic screens for C. elegans mutants with altered sensitivity to bacterial infections of the cuticle have recovered many mutants with associated chemical hypersensitivity and cuticle fragility phenotypes, presumably due to changes in the cuticle or surface properties of the worm (Gravato-Nobre et al. 2005). Among these hypersensitive mutants are those that fail to show the characteristic response to Microbacterium nematophilum infection, a swollen rectal epidermis; such mutants have the “Bus” phenotype (bacterially unswollen). Interestingly, mutations in certain bus genes confer greater sensitivity to cuticle infection by bacterial species to which wild-type worms are resistant. The bacterial sensitivity of bus mutants can be suppressed by mutation in suppressor of bus, or subs genes (Hodgkin et al. 2013, see Materials and Methods). Among subs mutants, we found eight independent mutant alleles of agmo-1, the C. elegans ortholog of the human alkylglycerol monooxygenase (AGMO) gene. Similar to cat-4 and ptps-1 mutants, agmo-1 mutants displayed rapid cuticle disintegration in standard alkaline bleach (Figure 4A). Four agmo-1 alleles contain premature stop codons, so are likely complete loss of function (Figure 6). Four missense alleles alter amino acids highly conserved among putative AGMO orthologs from other metazoans (Figure 6B, Figure S6), so may abolish or greatly reduce protein function; two of these affect conserved histidines that are among the eight conserved histidines characteristic of fatty acid hydroxylases.

Figure 6.

Mutations altering pathogen sensitivity and cuticle strength map to agmo-1. (A) Physical map of agmo-1 region with gene model and location of agmo-1 mutant alleles (arrows). All alleles were isolated as suppressors of lethality to Leucobacter Verde1 in bus or srf mutants. Sequencing of C. elegans agmo-1 cDNAs confirms the gene model shown (Figure S6), and the predicted protein used for alignments. Image is partly derived from the WormBase genome browser. (B) Alignments of predicted C. elegans AGMO protein with human AGMO and putative AGMO proteins from other metazoans. Sequences selected for alignment were best BLASTP matches with both C. elegans and human AGMO proteins; C. elegans AGMO-1 is ∼38% identical to human AGMO. Most notations are as in Figure 2. Eight 100%-conserved histidines (red) are found in all fatty acid hydroxylases and are required for human AGMO function (Watschinger et al. 2010). A conserved Glu (E154) likely required for biopterin binding is indicated in green; additional residues essential for human AGMO activity are in purple (Watschinger et al. 2012). Species abbreviations are as in Figure 2, except Tca, Tribolium castaneum and Xtr, Xenopus tropicalis.

Two new mutant alleles of cat-4 and one ptps-1 allele were also isolated via this screen, including a hypomorphic/reduction-of-function allele cat-4(e3015), indicating that complete loss of biopterin synthesis is not required for the Subs phenotype. All agmo-1 mutants tested were hypersensitive to a variety of chemicals, including alkaline bleach, SDS, and levamisole, comparable to cat-4 and ptps-1 mutants (Figure 4C, Figure S5, B and D). As expected, agmo-1 mutants displayed normal levels of DA and 5HT in neurons (Figure S7), indicating the independence of the neuronal and epidermal roles of BH4.

The BH4 regeneration cycle maintains biogenic amine levels under conditions of limiting BH4

We further tested putative knockout alleles of the worm biopterin regeneration enzyme genes qdpr-1 and pcbd-1. These mutants showed no obvious morphological or behavioral phenotypes, and initially appeared wild type for 5HT, DA, and bleach sensitivity (Figure 3, Table 2). Because loss of BH4 regeneration in the context of normal de novo BH4 synthesis may not reduce BH4 levels sufficiently to cause an obvious phenotype, we also tested qdpr-1 and pcbd-1 mutants in combination with the cat-4 reduction-of-function allele, cat-4(e3015). Reduction of neurotransmitters in cat-4(e3015) worms was most apparent in young larvae; e3015 adult worms were nearly wild type, suggesting that functional BH4 accumulates over the life of the worm (Figure S9). We found that both 5HT and DA were very strongly reduced in double mutants with qdpr-1 and pcbd-1 in comparison to the single mutant cat-4(e3015). The double mutants also showed the strongest reduction in young larvae (Figure 7), although differences were also apparent among adults (Figure S10). For example, although 5HT-IR was absent in only 7–20% of L1–L2 cat-4 worms, 55% of pcbd-1(tm5924); cat-4(e3015) worms lacked 5HT staining (Figure 7, A and C). Similarly, 82% of qdpr-1(tm2373); cat-4(e3015) and 88% of qdpr-1(tm2337); cat-4(e3015) lacked 5HT staining (Figure 7, B and D). Similar results were obtained using FIF staining to detect DA (Figure S11). These results demonstrate an important role for biopterin regeneration under conditions in which biopterin levels may be limiting for neurotransmitter synthesis.

Figure 7.

5HT synthesis is dependent on the biopterin regeneration pathway. Serotonin immunoreactivity (5HT-IR) in NSM somas and neurites of young (L1–L2) worms. Mixed populations of wild type (N2), single, and double mutant worms were scored by serotonin antibody staining. Staining definitions are as follows: strong, somas bright (saturated staining, no internal features apparent) and neurites bright; medium, somas not saturated, may show some internal structure (i.e., a nucleus may be seen), neurites present; weak/faint, somas visible (may be just above background), neurites absent or very faint; and none, no stained structures apparent. (A) pcbd-1(tm5924) mutants display at most a mild reduction in 5HT-IR. Double mutant pcbd-1; cat-4(rof) worms display almost no 5HT-IR; cat-4(rof) = reduction of function allele = e3015. (B) Loss of 5HT-IR in qdpr-1 mutant worms. Wild type (N2), qdpr-1, cat-4, and double mutant qdpr-1; cat-4 worms scored as described in A. qdpr-1(a) = tm2337; qdpr-1(b) = tm2373. (C) Examples of staining in mutants including pcbd-1(tm5924). (D) Examples of staining in mutants including qdpr-1(tm2337). All images in C and D were taken using the same exposure.

BH4 synthesis genes are expressed in biogenic amine neurons and in the epidermis

Consistent with the known role of GTPCH1 in synthesizing BH4, required for the function of AAAHs, GFP reporter constructs show cat-4 expression in identified serotonergic and dopaminergic neurons (where TPH and TH are expressed, respectively), and in the epidermis (where both PAH and AGMO are expressed, see below). We examined cat-4 reporter constructs from three different sources, including some previously described by others (see Materials and Methods; Table S1 and Table S2). Overall, in larval and adult worms, we observed strong expression in serotonergic and dopaminergic neurons, in most of the epidermis—especially the large epidermal syncytium (hyp7)—and more weakly in some intestinal cells (Figure 8A). Given the phenotypes of ptps-1 mutants, and the predicted function of the gene, we expected that the gene’s expression would likely match that of cat-4/GTPCH1. We made and examined three independent ptps-1::GFP reporter transgenics (translational fusion in the final coding exon plus 2.0 kb of sequence upstream of the predicted ATG), as well as a ptps-1 reporter transgene (∼2.6 kb upstream sequence) described elsewhere (Zhang et al. 2014). Transgenics with the shorter upstream sequence resembled the cat-4 expression pattern, with expression in some epidermal cells, and a few serotonergic neurons (Figure 8B). The ptps-1 reporters, however, were also observed in additional cells, and were not expressed in other epidermal cells or in all 5HT and DA neurons, suggesting these reporters may lack important positive or negative regulatory elements.

Figure 8.

Biopterin synthetic genes are expressed in biogenic amine neurons and in the epidermis. (A) Expression of cat-4 GFP reporters in larval stage 2 (L2) worms, construct with ∼2.7 kb upstream sequence (strain CZ9719). Top: Superficial focal plane showing epidermal expression, especially in the hyp7 syncytium. Seam cells have undergone doubling division and can be seen along the lateral side as darker regions among the brightly staining hyp7 cell. Dendritic endings of CEP neurons can be seen at the tip of the “nose.” Middle: Medial focal plane showing anal cells, strongly expressing tail epidermal cells and posterior intestinal cells expressing GFP. Bottom: Medial focal plane showing epidermal expression in the body and head. In these transgenics, expression in the head is seen in hyp6, but not the more anterior hyp5. NSM and CEP neuron somas are seen in the head, plus some neuronal processes (especially CEP processes). A few other neuronal somas stain less brightly. The anterior intestine also shows GFP expression, as do some rectal epithelial cells (likely B and Y cells). (B) ptps-1-GFP reporters. Left: head showing NSM expression plus other neurons and nonneuronal cells, including anterior intestine (dorsal–ventral view, maximum intensity projection) (MIP); strain OH11619, described by Zhang et al. (2014). Right: vulval region showing expression in neurons VC4, 5 (weak) and HSN, plus lateral seam cells (lateral view, MIP); strain CZ18321. (C) pcbd-1 reporter in larva showing expression in epidermal syncytium hyp7, tail and head epidermis, but not in seam cells; strain CZ19212. Out of focus: pharyngeal muscle and rectal epithelium expression. Inset: enlargement of midbody region. (D) qdpr-1 reporters. Top images: single optical sections from larva showing broad epidermal expression, but little expression in neurons; strain CZ19213. Insets: enlargements of midbody regions. Bottom image: Head showing expression in identified 5HT and DA neurons: NSMs, ADFs, CEPs, and other cells, with reduced epidermal expression; strain CZ19215 (lateral view, MIP). (E) gfrp-1 reporter showing expression in epidermal syncytium hyp7 (top: superficial focal plane) and other epidermal cells (bottom: central focal plane). (A–E) Anterior is to the left in all worms. (A, C, and E) Standard epifluorescence. (B and D) Laser scanning confocal imaging; (B and D, bottom) images are maximal intensity projection of Z-stack; and (D, top and middle) single confocal image planes.

We made and examined reporter transgenics for the pcbd-1, qdpr-1, and gfrp-1 genes, and also examined some described by others (Zhang et al. 2014). All were expressed in the epidermis, similar to that observed in cat-4 transgenics (Figure 8, C–E), but were not highly expressed in 5HT and DA neurons. The epidermal expression, when observed, was similar in localization, intensity, and developmental timing in larvae. All of these transgenics also showed some expression of varying intensity in other cells (nonepidermal cells, and non-5HT and non-DA neurons in the head and body). When we examined one qdpr-1 transgenic with 4.9-kb genomic sequence (full-length translational fusion with ∼4.0 kb upstream), we observed expression also in several known 5HT and DA neurons (Figure 8D, bottom panel). This transgenic showed some reduction in epidermal expression. With the exception of the larger qdpr-1 construct, the other previously described transgenics (Zhang et al. 2014) were expressed similarly to those reported here.

agmo-1 is expressed throughout the epidermis beginning in embryogenesis

We examined agmo-1::GFP reporter fusion transgenics and found that an agmo-1 transgene with ∼2300 bp of upstream sequence was expressed in the epidermis similarly, but not identical to cat-4 reporter constructs (Figure 9). agmo-1::GFP reporter constructs were expressed more broadly in the epidermis than were cat-4 or pah-1 reporter constructs (Figure 9, A–E, Figure S12). We typically saw expression in all epidermal syncytia and in seam cells at all stages, beginning in the embryo. Based on GFP intensity, agmo-1 is particularly highly expressed in late stage embryos when the L1 cuticle is being constructed prior to hatching. cat-4 and pah-1 reporters are also expressed in embryos, although expression appears earlier in embryogenesis (Figure S13).

Figure 9.

agmo-1 is expressed in the C. elegans epidermis. (A–G) Expression of agmo-1::GFP reporter fusions in the epidermis; all worms at same magnification, standard epifluorescence. (A–E) Young adult hermaphrodites; anterior to the left. (A) Head, surface focal plane showing expression in epidermal syncytial cells and seam. (B) Head, mid-depth focal plane showing expression in all head epidermal cells (same head as A). (C) Lateral body surface showing expression in epidermal syncytium hyp7 and lateral seam cells. (D) Mid-depth focal plan showing expression in vulval cells and ventral epidermal ridge. (E) Tail showing strong expression in rectal epithelial cells and other tail epidermis. (F) Two 3-fold-stage embryos with very strong expression in epidermis. (G) Young larva, mid-depth focal plane, showing expression like that seen in adult (and all other larval stages).

Discussion

We have characterized the genetic basis of biopterin/BH4 biosynthesis and function in C. elegans. We have demonstrated that the mutationally identified cat-4 gene encodes GTPCH1, and that mutations in a PTPS-encoding gene cause phenotypes identical to those of cat-4 mutants. cat-4 and ptps-1 mutants lack their respective enzymatic activities, have greatly decreased BH4 content, and have phenotypes consistent with their predicted roles in BH4 synthesis, including loss of the neurotransmitters 5HT and DA. Biopterin synthesis mutants were not rescued by supplemental BH4 or sepiapterin. Although some exogenous biopterin was taken up, we could only detect the inactive form 7,8-dihydrobiopterin (BH2). In mammalian cell culture and in mice, BH4 is taken up as BH2 and then reduced to BH4 by DHFR (Hasegawa et al. 2005). In C. elegans, we hypothesize that exogenous BH4 is similarly taken up as BH2, but for unknown reasons is not subsequently reduced to BH4 by the C. elegans DHFR. In any case, the lack of functional rescue of the BH4-deficient mutants by exogenous biopterins is consistent with the general requirement for biopterin synthesis in cells and tissues that use BH4-dependent enzymes (Thöny et al. 2000).

Our observations, using 5HT and DA synthesis as a proxy for biopterin levels, demonstrate that pcbd-1 and qdpr-1 genes, predicted to encode BH4 regeneration enzymes, function to maintain biopterin needed for neurotransmitter synthesis. The biopterin regeneration mutants showed clear effects on neurotransmitter synthesis when combined with a reduction-of-function cat-4 mutation. Because of the inherent variability of both techniques—immunofluorescence (for 5HT) and FIF (for DA)—possible slight reductions in staining observed in the pcbd-1 or qdpr-1 single mutants could not be definitively assessed. Thus in C. elegans, function of these BH4 regeneration genes may be specifically revealed when BH4 biosynthesis is impaired.

Besides lacking 5HT and DA, cat-4 and ptps-1 mutants display increased sensitivity to many chemicals, and a less sturdy cuticle. The cuticle defects of cat-4 have been a long-standing conundrum, as they are not seen in mutants in known BH4-dependent AAAH enzymes (pah-1, cat-2, and tph-1). Solution of this puzzle awaited the cloning of the last remaining BH4-dependent enzyme, alkylglycerol monooxygenase/AGMO (Watschinger et al. 2010), allowing identification of the C. elegans AGMO ortholog. Mutations in agmo-1 were predicted to cause cuticle fragility and hypersensitivity, and such mutations were indeed subsequently found in screens for altered resistance to a worm bacterial pathogen that infects animals via the cuticle (Hodgkin et al. 2013). Consistent with a role in establishing outer surface properties of the worm, agmo-1 reporters are expressed throughout the epidermis, which secretes the cuticle, a complex extracellular matrix (Chisholm and Xu 2012). The epidermis is also most likely the cellular origin of the epicuticle, a poorly understood extracellular lipid-rich layer that has been hypothesized to function as a permeability barrier and in pathogen defense (Chisholm and Xu 2012). Identification of AGMO as required for cuticle integrity further supports the idea that lipid metabolism is critical for surface properties of nematodes.

Our analysis of biopterin-related gene expression in C. elegans by reporter fusion transgenics is generally consistent with BH4 synthesis being required cell autonomously for AAAH function in 5HT and DA neurons and for AGMO function in the epidermis. Not all our reporter transgenics, however, were expressed in patterns expected based on mutant phenotypes of the genes examined. Transgenes reported here are mostly transcriptional fusions and may lack positive or negative regulatory elements. Further work, using rescuing transgenes or under endogenous or tissue-specific control, will be required to fully define the cellular requirements for BH4 synthesis.

Our studies reveal the first in vivo biological role for the BH4-dependent lipid metabolic enzyme AGMO, which is the only enzyme known to degrade the ether lipid bond in alkylglycerols and alkylglycerol-lyso-phospholipids (Watschinger and Werner 2013). The precise biochemical role of AGMO in the cuticle and permeability barrier in the worm remains to be established. We do not know whether alkyl ether lipid metabolism in C. elegans serves primarily an anabolic or catabolic function: agmo-1 mutant phenotypes could result from failure to synthesize a needed product or by accumulation of toxic intermediates. Failure of the AGMO reaction might lead to accumulation of ether lipids, somehow destabilizing the lipid-rich epicuticle. Alternatively, AGMO-1 substrates might alter cuticle development via signaling pathways, analogous to the effects of antitumor ether lipids on mammalian tumor cells (Arthur and Bittman 1998).

Other mutants with altered pathogen sensitivity in C. elegans display chemical hypersensitivity and cuticle fragility. The bus mutants are partially resistant to bacterial rectal infection by certain worm pathogens, apparently due to alterations in cuticle surface features (Gravato-Nobre et al. 2005). Some of these mutants are bleach hypersensitive, dying and rupturing more quickly than N2 wild type. Several bus genes have now been identified, and some encode lipid metabolic enzymes expressed in cells that overlap with cat-4 expression, suggesting a pathway association with agmo-1. For example, bus-18 (aka acl-10) encodes a lysocardiolipin acyltransferase expressed in the epidermis (Gravato-Nobre and Hodgkin 2008; Imae et al. 2010). However, in general, the cuticle defects of such bus mutants are more severe than those of biopterin synthesis or agmo-1 mutants and extend to outer cuticle sloughing during handling, rupture at the vulva, and “skiddy” locomotion. These more severe cuticle defects suggest additional, widespread roles for lipid metabolism in the cuticle.

In conclusion, our systematic analysis of BH4 function in C. elegans confirms its essential role as a cofactor for TH and TPH in neurotransmitter synthesis in neurons. In addition to this well-established function, we elucidate here a new role for BH4 as a cofactor for AGMO in epidermal cells. Our studies reveal an unexpected in vivo role for AGMO in supporting cuticle stability and sensitivity to bacterial infection, presumably via ether lipid metabolism. These observations raise the question of whether AGMO might play comparable roles in epidermal function in other animals or in humans. The presence of a complete enzymatic pathway for de novo synthesis of BH4 has been established in the human epidermis (Chavan et al. 2006), but at present no mutations in mammalian or human AGMO have been reported, nor have severe skin defects been described in patients deficient in BH4 synthesis. Nevertheless, we note that the products of AGMO catalysis are converted to fatty acids by fatty aldehyde dehydrogenase (FALDH), and that FALDH deficiency in humans results in a combination of skin permeability barrier defects, ichthyosis, and neurological disease known as Sjögren-Larsson syndrome (Rizzo 2011). It would be interesting to explore whether AGMO contributes to lipid-based permeability barrier function in other organisms or in humans.