Lysosomal reduced thiols are essential for mouse embryonic development

Significance

The study of the lysosomal cysteine import system was limited until the discovery of a key component, the transmembrane protein MFSD12. Here, we genetically disrupted Mfsd12 in mice and found that it is essential for organogenesis. Parallel disruption of the gene for cystinosin (CTNS), which exports cystine (oxidized cysteine) from lysosomes, did not elicit similar defects, suggesting that the essential function of MFSD12 is to provide reduced thiols in the form of cysteine to lysosomes. Treatment of breeding mice with cysteamine, an analog of cysteine that does not require lysosomal transport, rescued the development of Mfsd12-deficient mice. Our results suggest that lysosomal thiols are essential for animal development and highlight the role of cysteine as a reducing agent in lysosomes.

Abstract

While it has been appreciated for decades that lysosomes can import cysteine, its significance for whole-organism physiology has remained uncertain. Recent work identified MFSD12 as a transmembrane protein required for cysteine import into lysosomes (and melanosomes), enabling genetic interrogation of this pathway. Here, we show that Mfsd12 knockout mice die between embryonic days 10.5 and 12.5, indicating that MFSD12 is essential for organogenesis. Mfsd12 loss results in the expression of genes involved in cellular stress and thiol metabolism and likely disproportionately affects the erythroid, myeloid, and neuronal lineages. Within lysosomes, imported cysteine is largely oxidized to cystine, which is exported to the cytosol by the cystinosin (CTNS) transporter. However, unlike Mfsd12, loss of Ctns is compatible with viability, suggesting that the essential role of MFSD12 lies not in supplying cystine to the cytosol, but in providing reduced cysteine within the lysosomal lumen. Supporting this model, maternal treatment with cysteamine—a lysosome-penetrant thiol—rescued the development of Mfsd12 knockout embryos, yielding viable adult offspring. These findings establish lysosomal thiol import as a critical metabolic pathway and provide genetic tools to further clarify its physiological and biochemical roles.

Proteins, nucleic acids, carbohydrates, and lipids are trafficked to the lysosome and broken down into their constitutive parts for reuse (1, 2) by the over 60 enzymes that operate within the carefully controlled environment of its lumen. A key metabolic feature of the lysosome is its thiol redox status. While the lysosome is overall oxidizing compared to the cytosol, proteins destined for degradation often have disulfide bonds that must be reduced to facilitate proteolysis, and the cysteine cathepsin proteases must maintain a reduced cysteine in their active sites to be functional (3–7). Glutathione, the dominant reductive agent in the cytosol, does not appear to be transported by lysosomes (5, 6). Instead, lysosomes utilize cysteine for thiol reduction, importing it from the cytosol in a reduced form and exporting it as cystine after its oxidation (8–11). Two separate lysosomal transport systems control cysteine influx and cystine efflux, and together establish a pathway whose activity has been well described in cell-free systems (9, 11). However, until recently the specific component(s) that comprise the cysteine import system were unknown, so it was not possible to genetically manipulate cysteine import to evaluate its significance in vivo.

While studying the human pigmentation gene MFSD12, we recently discovered that MFSD12, which is a 12-transmembrane domain protein, is required to import cysteine into both melanosomes and the lysosomes of unpigmented cells (10). It is still unknown if the MFSD12 protein can directly transport cysteine when reconstituted into liposomes, but its function in cysteine transport appears quite central. Isolated lysosomes from MFSD12 knockout cells have defective cysteine uptake, and relocalization of the MFSD12 protein to the plasma membrane is sufficient to introduce a new cell surface cysteine transport activity (10, 12). These findings establish MFSD12 as a promising genetic target to study the organismal importance of lysosomal cysteine.

Insights into the potential in vivo significance of lysosomal cysteine transport come from studies of cystinosin (CTNS), the well-established lysosomal cystine exporter (11, 13). Mutations in the CTNS gene cause the inherited disease cystinosis, which manifests in renal Fanconi syndrome, failure to thrive, and the progressive destruction of additional organ systems, which is thought to result from the toxic accumulation of cystine (13, 14). Recent studies examining the role of CTNS in models of development and cancer suggest that lysosomal cystine buffers cytosolic cysteine during oxidative stress and can influence pro-growth signaling (15–20). However, because the clinical features of cystinosis are thought to be driven in large part by mechanical damage from cystine crystals (13), it is challenging to interpret the disease phenotypes as purely caused by a lack of cystine export. Here, by comparing Mfsd12 and Ctns knockout mice and manipulating lysosomal thiol content with cysteamine, we reveal lysosomal thiols as essential for embryonic development.

Results

Generation of an Mfsd12 Knockout Allele in Mice.

We generated an Mfsd12 mutant allele using CRISPR-Cas9 and an sgRNA designed to target all known isoforms of Mfsd12 (Fig. 1A). The resulting 2-bp insertion (c.360_361insGA or Mfsd12^em1Hadel) disrupts the fourth of the 12 predicted transmembrane domains of MFSD12 before generating a stop codon 28 amino acids downstream (Fig. 1B, Uniprot Q3U481). Because we initially intended to examine potential pigmentation phenotypes in these mice, we backcrossed them with congenic C57BL/6J-A^w-J mice and bred white-bellied Jackson agouti (A^w-J, red-brown and black fur) in place of nonagouti (a, black fur) (21). Consistent with the action of nonsense-mediated decay, Mfsd12 mRNA expression was decreased to ~60% of wild-type levels in kidney and spleen extracts from Mfsd12^em1Hadel heterozygous mice (Fig. 1C).

Fig. 1.

Loss of Mfsd12 in mice results in completely penetrant lethality. (A) Diagram of murine Mfsd12 gene structure and the location of the CRISPR-induced Mfsd12 mutation. (B) Nucleotide and amino acid resolution diagram of Mfsd12^em1Hadel mutation, showing the predicted amino acid sequence of the Mfsd12 knockout allele. This includes an annotation of transmembrane domain IV from Uniprot Q3U481. Note the loss of hydrophobic amino acids (I, V, F etc.) (C) qRT-PCR analysis of Mfsd12 expression in the kidney and spleen comparing Mfsd12^+/+ and Mfsd12^+/− matched, male littermate pairs. Data are shown as mean ± SEM, with significance determined by unpaired, two-tailed Student’s t test (*P < 0.05, **P < 0.01, n = 3 pairs of mice). (D) Offspring genotypes from 12 monogamous Mfsd12^+/− x Mfsd12^+/− crosses. The P-values are from a Chi-squared test for Mendelian ratios (n = 123 mice). (E) Representative photographs of 4-wk-old Mfsd12^+/+ and Mfsd12^+/− littermates show no obvious differences in body size or coat pigmentation.

We analyzed 12 separate monogamous crosses of Mfsd12 heterozygous mice and, despite weaning 123 mice at 3 to 4 wk of age, did not isolate a single Mfsd12 knockout animal (Fig. 1D). Heterozygous animals appeared at a rate ~1.9-fold that of wild-type animals, suggesting loss of one copy of Mfsd12 did not cause an intermediate viability phenotype. This is consistent with Mfsd12 heterozygous mice having no detectable differences compared to wildtype counterparts in fur pigmentation or 4-wk postnatal body weight (Fig. 1E and SI Appendix, Fig. S1A).

Mfsd12 Knockout Mice Show Development Defects at 10.5 to 12.5 d Post Fertilization.

To better understand why Mfsd12 is essential, we sought to identify the stage at which the Mfsd12 knockout embryos cease to develop. Timed matings were set up with heterozygous Mfsd12 mice, and embryos were isolated at defined time points before precise embryonic timing (“mE”) was further refined with development landmarks and computational analysis of limb morphology (22, 23). Knockout embryos isolated between embryonic days 10.5 and 12.5 were grossly abnormal (10 out of 10), with defects becoming most profound on the 12th day of embryonic development (Fig. 2 A–D). Early-stage defects (mE10.5-11.5) were typically small size and reduced fetal blood (Fig. 2A), while later-stage defects (mE12.0-13.0) included underdeveloped limbs and facial features (Fig. 2B). We interpreted these latter Mfsd12 knockout embryos to be essentially dead. Consistent with this interpretation, very few Mfsd12 knockout embryos were isolated after mE12.5, and none exhibited normal morphology (Fig. 2 C and D).

Fig. 2.

Mfsd12 knockout induced embryonic lethality manifests 10.5 to 12.5 d after fertilization. (A and B) Photographs of littermate embryos staged at mE11.0 and mE12.2 illustrate the progressive defects in Mfsd12 knockout embryos. All Mfsd12 knockout embryos shown were classified as “abnormal” based on size and morphology. Precise embryo staging was done by averaging limb staging data from eMOSS tool for “normal embryos” per litter (22, 23). (C and D) Tables showing the number of embryos isolated at indicated days post-fertilization and their genotypes. Embryos that were not visibly different from wild type are included in Fig. 2C, while those that were deemed grossly abnormal based on morphology or small size are in Fig. 2D. The P-values are from a Chi-squared test for Mendelian ratios calculated for “Normal Embryos” (n = 61, 42, 40 embryos). (E) Volcano plot of differential gene expression analysis comparing the protein-coding, autosomal transcriptomes of Mfsd12 knockout and wild-type embryos (n = 4 per genotype). (F and G) GSEA of Mfsd12 knockout versus wild-type embryos using the Hallmark pathways collection (via MSigDB) and the Mouse Cell Atlas (24, 25). The top 10 gene categories by FDR are plotted for each analysis. The legend is shared between panels. (H) ChEA3 transcription enrichment using differentially expressed genes from Mfsd12 knockout versus wild-type embryos. The top 10 transcription factors identified by mean rank are plotted.

Mfsd12-Related Developmental Defects Include an Embryo-Wide Transcriptional Stress Response.

To begin to address the molecular basis of why Mfsd12 is an essential gene, we performed gene expression profiling on whole embryos staged between mE10.0-11.0 d. Transcriptome comparisons between wild-type and knockout embryos identified substantial gene expression changes, including strong downregulation of Mfsd12 itself (Fig. 2E and Dataset S1). Upregulated transcripts included markers of cell-stress responses (Chac1, Sesn2, Atf3, Asns, Eif4ebp1, and Trib3) and genes directly tied to thiol metabolism (Chac1, Slc7a11, and Cth) (26–38). Gene set enrichment analysis (GSEA) of the Hallmark pathways (MSigDB) identified the “Unfolded Protein Response” as the top enriched program and “Heme Metabolism” as significantly downregulated (Fig. 2F) (24, 39). In line with this latter finding, enrichment analysis for lineage-specific markers nominated erythrocytes, neurons, and myeloid lineages as selectively depleted in Mfsd12 knockout embryos (Fig. 2G) (25, 39). ChEA3 analysis implicated DDIT3/CHOP and ATF3 (known regulators of cell-stress responses) and NFE2, KLF1, and GATA1 (known drivers of erythropoiesis) as potential drivers of the transcriptional response in Mfsd12 knockout embryos (Fig. 2H) (31, 40–45).

No significant transcriptional differences were observed between wild-type and heterozygous embryos, though a focused analysis confirmed a ~twofold reduction in Mfsd12 mRNA expression in heterozygous embryos (SI Appendix, Fig. S1 B and C and Dataset S1). In parallel, we performed thiol-focused chemoproteomic profiling and found no consistent differences between wild-type and heterozygous animals, even in lysosomal proteins (Dataset S2). While the consequences of total Mfsd12 loss are profound, its expression is by no means limiting, as a twofold reduction in Mfsd12 (in heterozygous mice) induces no phenotype even when assessed by sensitive molecular analyses.

The Knockout of Ctns Does Not Phenocopy That of Mfsd12 Knockout.

MFSD12 and CTNS act on two ends of a pathway, controlling cysteine and cystine movement through lysosomes (9, 10, 46). We hypothesized that by comparing loss-of-function phenotypes for both genes we would understand how lysosomal cysteine is important for development. Several Ctns knockout alleles already exist, but we opted to generate a one in parallel with Mfsd12 to control for any confounding factors in genetic backgrounds and husbandry environments, as the former has been reported to affect both Mfsd12 and Ctns phenotypes (14, 47).

Our Ctns-targeting sgRNA introduced a 6-bp deletion compounded with a 1-bp insertion (Fig. 3A, c.329_335delinsT, Ctns^em1Hadel). This alteration occurs upstream of the part of the gene that encodes the first transmembrane domain of the seven-transmembrane CTNS protein and leads to a premature stop codon 12 amino acids downstream (Fig. 3B, Uniprot P57757). Consistent with nonsense-mediated decay, we detected a decrease in Ctns levels to ~40 to 60% of wild-type levels in kidney and spleen from Ctns^−/− mice (Fig. 3C).

Fig. 3.

Knockout of Ctns does not phenocopy that of Mfsd12 with respect to embryonic lethality. (A) Diagram of Ctns gene structure and the location of the CRISPR-induced Ctns mutation. The diagram is oriented in the antisense direction. (B) Nucleotide and amino acid resolution diagram of Ctns^em1Hadel mutation. (C) qRT-PCR analysis of Ctns expression comparing three Ctns^+/+ and Ctns^−/− matched male littermate pairs. Data are shown as mean ± SEM, with significance determined by unpaired, two-tailed Student’s t test (****P < 0.0001, n = 3 pairs of mice). (D) Number and frequency from 4 monogamous Ctns^+/− x Ctns^+/− crosses. The P-values are from a Chi-squared test for Mendelian ratios (n = 89 mice). (E) Photograph of 4-wk-old Ctns^+/+ versus Ctns^−/− littermates showing no obvious coat color differences.

After weaning 89 mice from 4 monogamous Ctns^+/− crosses, we isolated several live Ctns knockout mice at 3 to 4 wk of age. These knockout animals were isolated at a lower-than expected frequency (10% versus 25%; see Fig. 3D) and had reduced body weights (SI Appendix, Fig. S2A). However, they were viable and fertile and were used to breed further homozygous Ctns knockout mice (SI Appendix, Fig. S2B). No pigment defects were observed in Ctns knockout animals, contrary to findings in other mouse models and in patients with cystinosis (Fig. 4D) (48).

Fig. 4.

Lysosomal thiol supplementation with cysteamine rescues Mfsd12 knockout animals to live birth. (A) Genotype ratios of neonatal mice (found both deceased and alive) identified on the day they were born (P0) from mothers treated with control water (0.3% sucrose, pH 6.0), cysteine-supplemented water (8 mM cysteine, 0.3% sucrose, pH 6.0) or cysteamine-supplemented water (8 mM cysteamine, 0.3% sucrose, pH 6.0). The P-values are from a Chi-square test for Mendelian ratios (n = 58, 35, 99 neonates) (B) Photograph of anesthetized neonates including live and deceased Mfsd12^−/− neonates, typical of cysteamine rescue. For each treatment category, littermates are shown. Photographs of control and cysteamine-treated mice were obtained on different days. (C) Neonate forepaws (ventral right and left paws) from the same treatment as (A and B). The paws feature digits parallel to the midline and the presence of distal phalanges, features of development associated with progression past ~E16 (23). (D and E) Schematic and results from an experiment testing the viability of Mfsd12 knockout mice after withdrawal of cysteamine at postnatal day 7. The P-values are from a Chi-square test for Mendelian ratios (n = 34 neonates). (F) Photograph of 3-mo old adult female littermates which were withdrawn from cysteamine 1-wk postnatally. The gray coat color of the Mfsd12^−/− animal is consistent with other reports of Mfsd12 deficiency on an agouti background (47, 49). (G) Model of lysosomal cysteine and cystine metabolism in mouse embryonic development, highlighting the importance of luminal thiols like cysteine as reducing agents over cystine storage and release to the cytosol.

Efforts to breed Mfsd12^+/−; Ctns^−/− mice to test for a direct genetic interaction between Mfsd12 and Ctns were hindered by decreased offspring numbers (2.4 mice per litter measured over 5 litters from 3 breeding pairs) and the death of offspring before 3 wk of age (10 out of 12 mice).

Cysteamine Treatment Rescues the Embryonic Development of Mfsd12 Knockout Mice to Live Birth.

Our observation that the knockout of Mfsd12 is lethal in mice whereas that of Ctns is not suggests that the essential function of MFSD12 in development is not to enable cystine storage for use in the cytosol via transport out of lysosomes by CTNS. We reasoned that the Mfsd12 knockout phenotype could be caused instead by the lack of reduced thiol import, a pathway shown in vitro to be important for degrading disulfide-bond rich substrates in lysosomes (3–7).

To test this hypothesis, we treated Mfsd12 heterozygous breeding pairs with cysteamine in their drinking water. Cysteamine is used in cystinosis therapy to convert cystine to cystine-cysteamine disulfides, which are effluxed from the lysosome through a different transporter (50, 51). Cysteamine is highly bioavailable and well-documented to enter lysosomes in a transport-independent manner (13). Consistent with our embryonic studies, no neonates (P0) were isolated from Mfsd12^+/− x Mfsd12^+/- breeding mice treated with a control drinking water (Fig. 4 A and B, n = 58 mice). However, when we supplemented the water with 8 mM cysteamine, knockout neonates were isolated at a frequency of 15% and at genotypic ratios that were overall consistent with Mendelian transmission (Fig. 4 A and B, P = 0.059, Chi-square test for Mendelian transmission). As a control for adding free-thiol antioxidants to drinking water, we also treated breeding mice with 8 mM cysteine, which did not lead to the appearance of live Mfsd12 knockout neonates (Fig. 4A, n = 35 mice).

Mfsd12 knockout neonates born to cysteamine-treated mothers had features indicating that they had advanced in development well past the stage at which the knockouts perished in the absence of cysteamine (Figs. 2 A and B and 4A). These included well-developed eyes, skin wrinkling, and erupted distal phalanges that had converged to the midline of the forepaw (Fig. 4 B and C and SI Appendix, Fig. S3A). Skin development had also progressed to stages consistent with late embryogenesis, including cornification, panniculus carnosus formation, and hair follicle development (SI Appendix, Fig. S3A) (52, 53).

Our cysteamine rescue was partial. The Mfsd12 knockout neonates weighed less than their wild-type and heterozygous counterparts from the cysteamine-treated crosses (SI Appendix, Fig. S3B). Of the Mfsd12 knockout neonates, 6 out of 15 were found dead. While this rate was significantly elevated compared to the rate of dead neonates found with control treatment (3 out of 58, P = 2.0 × 10⁻³, Fisher’s exact test), it was not significantly different from the rate observed in wild-type and heterozygous neonates with cysteamine treatment (18 out of 84, P = 0.19, Fisher’s exact test). This suggested that cysteamine itself adversely affected embryonic development, birth, or perhaps early mother-offspring nursing interactions. Cysteamine-treated mice had no obvious change in hair color (SI Appendix, Fig. S3C), suggesting that cysteamine, a known pigment-lightening agent (54, 55), might not have penetrated all tissues completely in our treatment regimen.

Discontinuation of cysteamine treatment in nursing mothers and neonates 1 wk after birth did not result in the death of Mfsd12 knockout mice (Fig. 4 D and E). 3 out of 34 mice weaned at 4-wk after birth were Mfsd12 knockouts (Fig. 4E), and these animals survived for months without supplemental cysteamine treatment, appearing grossly normal aside from a gunmetal gray fur color that was consistent with previous descriptions of mice with Mfsd12 deficiency (Fig. 4F) (47, 49).

Discussion

We find that in C57BL6/J-A^w-J mice, Mfsd12 is a completely penetrant essential gene. Mice lacking functional Mfsd12 die during the late stages of embryonic development associated with organogenesis. The essential role of MFSD12 is unlikely to be limited to facilitating cystine storage or the flux of cysteine/cystine through the lysosome, as loss of Ctns results in live, viable mice. Instead, its essential role likely involves providing cysteine as a lysosomal-reducing currency. Consistent with this, treatment of pregnant mice with cysteamine rescued the embryonic development of Mfsd12 knockout mice, and after birth, cysteamine was no longer required for the survival of Mfsd12 knockout mice. Thus, lysosomal cysteine transport is essential for animal embryonic development.

The defects we see in Mfsd12 knockout mice are similar to the embryonic defects that have been previously reported in zebrafish embryos upon morpholino-mediated mfsd12a disruption (46), the partial embryonic lethality of intestine-specific Mfsd12 knockout mice (12), and the embryonic lethality caused by whole-body loss of Mfsd12 in C57BL/6J mice, which was reported in the absence of primary data (47). Interestingly, viable Mfsd12 knockout mice have been reported in a mixed FVB/N x C57BL6/J background, and biallelic frameshift mutations in MFSD12 been reported in viable Shetland Ponies (47, 56). This suggests that outside variants might rescue the development of MFSD12-deficient animals. Further analyses of the interactions between Mfsd12 and other genetic loci might reveal new players in lysosomal cysteine metabolism.

Cysteamine treatment rescued the development of Mfsd12 knockout mice to produce live offspring, but the effect was incomplete. Most of the rescued Mfsd12 knockout neonates exhibited reduced body weight and died during or shortly after birth. This may reflect the limited efficacy of cysteamine, which is prone to oxidation in drinking water and has a short half-life in vivo—a known limitation in its use for treating cystinosis in humans (13). Furthermore, the high rate of death among wild-type and heterozygous neonates from cysteamine treatment suggests cysteamine-driven toxicities might synergize with the Mfsd12 knockout developmental defects. An important, and not mutually exclusive, implication is that a second mechanism may operate downstream of Mfsd12 and cannot be restored by cysteamine treatment. Recent work in cancer cell lines and in flies suggests that lysosomal cystine stores can replenish cytosolic cysteine, and buffer glutathione or acetyl-CoA during periods of oxidative stress (15–18, 20). The failure of the Ctns knockout to phenocopy that of Mfsd12 suggests a cystine storage mechanism is not the primary reason Mfsd12 is essential, but it does not rule out that it might play a contributing role.

The death of Mfsd12 embryos after the cellular stages of development suggests that the essential role of Mfsd12 is either cell type-specific or involved in the higher-order organization of tissues during organogenesis. The developmental defects we observed with germline Mfsd12 knockout were extensive and did not immediately suggest the dysfunction of a single tissue system or cell type; however transcriptomic analysis revealed several promising tissue lineages for further investigation (erythrocytes, neurons, and myeloid lineages).

We previously proposed that MFSD12 inhibition could be a therapeutic approach for cystinosis in humans (10), though studies in zebrafish have raised concerns about limited efficacy and dose-limiting developmental toxicity associated with mfsd12a disruption (26). Our finding that Mfsd12 is fully haplosufficient suggests that near-complete inhibition of MFSD12 may be necessary to achieve therapeutic benefit. However, the observation that Mfsd12 knockout mice can survive postnatally without cysteamine treatment indicates that the developmental toxicities observed with Mfsd12 loss do not reflect safety concerns for therapeutic MFSD12 inhibition in adults and children.

While we have viewed MFSD12 and CTNS as components of a linear pathway when considering cysteine and cystine metabolism, our findings suggest they do not function as a simple “bucket brigade” when considering their downstream effects. A deeper investigation into their distinct cellular and developmental roles in animal models will allow us to better delineate their relationship. In this work, we provide models and approaches to do exactly that.

Materials and Methods

Mouse Line Generation and Husbandry.

Mice studies and procedures were approved by the Institutional Animal Care and Use

Committees of the Whitehead Institute and Massachusetts General Hospital and were conducted strictly in accordance with the approved animal handling protocol. Animals were cared for according to the requirements of the National Research Council’s Guide for the Care and Use of Laboratory Animals.

C57BL/6J founder mice were generated with Cas9 and sgRNAs with the following targeting sequences:

Mfsd12: CGGGGAAGCCACGCCTGAAT

Ctns: TCACCAAGAACCGGATCCTG

After selecting frameshift alleles (Figs. 1A and 3A) for further study, mice were crossed to C57BL/6J-A^w-J (JAX#000051) to introduce the agouti coat color. Mice were genotyped using Sanger sequencing, next-generation amplicon sequencing (Quintara Biosciences), and a fluorescent probe-based qPCR assay implemented by TransnetYX interchangeably.

Mfsd12 and Ctns qRT-PCR.

Mouse organs from sibling pairs were harvested, cut into pieces, and stored in an RNA preservative solution (saturated ammonium sulfate, 20 mM EDTA, 25 mM sodium citrate, pH 5.2) before storage at −80 °C. Small pieces of tissue (~1 × 1 mm) were minced and placed in RLT Plus buffer (QIAgen, 1053393) supplemented with 140 mM 2-mercaptoethanol. Lysates were run-through QIAshredder columns (QIAgen, 79656) before RNA was isolated via QIAgen RNAeasy Plus kit (QIAgen, 74134) according to the manufacturer’s instructions.

Gene expression was measured using the KAPA SYBR One-Step qRT-PCR kit (Sigma, SF1UKB) using the manufacturer’s protocol and the following PCR primers:

Mfsd12_F: GAACGAACCCAATGAGCACACC

Mfsd12_R: CTGAGACAGGTTCACAATGAGCC

Ctns_F: ATGGCTCCAGTTCCTCTTCTGC

Ctns_R: AATCGAGGAGCACACCGCCAAT

Gapdh_F: CTCCCACTCTTCCACCTTCG

Gapdh_R: GCCTCTCTTGCTCAGTGTCC

Target gene C_T values were determined automatically in QuantStudio Design Analysis 2 software (Applied Biosystems), normalized to Gapdh, and presented as a fold-change value determined via the ∆∆C_T method.

Embryo Harvest, Staging, and Characterization.

Pregnant dams were sacrificed via CO₂-based euthanasia according to approved Institutional Animal Care and Use guidelines at the Massachusetts General Hospital at designated time points before harvest. Embryos were dissected in ice-cold PBS, and yolk-sacs or small tail sections were taken for genotyping before photo-documentation with a dissecting microscope. The precise embryonic day for each group of embryos was estimated with landmark features and with computational limb staging for embryos between mE10.25 and mE13.5 with eMOSS (https://limbstaging.embl.es) (22). For each litter of embryos, embryonic day values from “normal” embryos were averaged to estimate the reported age of the litter. Image scale bars were set by imaging a hemocytometer (Hausser Scientific) in parallel. Exposure conditions and brightness adjustments for each embryo image were not standardized.

Transcriptome Profiling and Analysis.

Whole embryos were mechanically homogenized with a handheld pestle (Kimble) and stored in Trizol solution (Invitrogen,15596026) at −80 °C. RNA was prepared by chloroform phase separation (0.2 mL chloroform in 1 mL Trizol), followed by isopropanol precipitation of the aqueous phase (0.5 mL) and washing of the resulting RNA pellet with 80% ethanol before drying and resuspending in nuclease-free water.

Total RNA was quantified using the Quant-iT RiboGreen RNA Assay Kit (Invitrogen) and normalized to 5 ng/μL. An aliquot of 325 ng of RNA from each sample was transferred into library preparation using an automated variant of the Illumina TruSeq Stranded mRNA Sample Preparation Kit. This method preserves strand orientation of the RNA transcript. It uses oligo dT beads to select mRNA from the total RNA sample and is followed by heat fragmentation and cDNA synthesis from the RNA template. After enrichment the libraries were quantified using Quant-iT PicoGreen (Invitrogen), samples were normalized to 5 ng/μL, and a library pool was assembled and quantified using the KAPA Library Quantification Kit (Roche). The entire process was performed in a 96-well format and all pipetting was done by either an Agilent Bravo or Hamilton Starlet instrument.

Pooled libraries were normalized to 2 nM and denatured using 0.1 N NaOH prior to sequencing. Flowcell cluster amplification and sequencing were performed according to the manufacturer’s protocols using the NovaSeq 6000 (Illumina). Each run was a 151 bp paired-end with an eight-base index barcode read. Data was analyzed using the DRAGEN Pipeline v4.2.4 which includes de-multiplexing, data aggregation, and alignment (GRCm39, NCBI accession: GCF_000001635.27).

Resulting BAM files were coordinate sorted. Gene-level quantification was performed using featureCounts from the Subread package, assigning uniquely mapped reads to annotated genes. Differential expression analysis was conducted using DESeq2. GSEA was carried out using WEB-based Gene SeT AnaLysis Toolkit (WebGestalt) (39), with signed significance scores (−log₁₀ adjusted P-value multiplied by the sign of the log₂‚ fold change) used as the ranking metric. Human orthologs of mouse genes were used for analysis with the Hallmark (MSigDB) and ChEA3 transcription factor target gene sets. Raw data and counts are available on GEO (GSE300056) (57).

Lysate Preparation for Chemoproteomic Treatments.

Whole embryos at mE12.0 to 12.5 (stored at −80 °C) were pulverized and lysed in DPBS supplemented with 0.05 U/µL benzonase (Santa Cruz) and 1× protease inhibitors (Roche, 4693132001) using a chilled bath sonicator (Q700, Qsonica). Lysates were centrifuged for 3 min at 300 × g at 4 °C. Protein concentrations were determined using a BCA assay (Thermo Fisher Scientific), and 50 µg of protein was used per compound treatment. Lysates were incubated at room temperature with vehicle (DMSO) or 500 µM KB03 (CAS No. 790-75-0, synthesized in house) for 1 h, followed by treatment with 1 mM desthiobiotin-iodoacetamide (DBIA, synthesized in house) for an additional hour.

Following DBIA incubation, lysates were reduced with 5 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP, Sigma-Aldrich) for 2 min at room temperature, then alkylated using 20 mM chloroacetamide (Sigma-Aldrich) for 30 min in the dark at room temperature. Proteins were precipitated using SP3 magnetic beads. Briefly, SP3 magnetic beads (Cytiva) were prewashed with LC-MS grade water (Sigma Aldrich), and 250 µg of combined SP3 beads (1:1, hydrophobic:hydrophilic) and LC-MS grade ethanol (Sigma Aldrich) were added to each sample to achieve a final concentration of 50% ethanol. SP3 incubation was performed for 30 min at room temperature, after which the beads were washed three times with 80% HPLC grade ethanol (Sigma Aldrich) and then resuspended with 175 µL of Trypsin/Lys-C (1 µg, Thermo Fisher Scientific) in 200 mM EPPS (Sigma Aldrich) pH 8.4, 5 mM CaCl₂. Proteins were digested overnight (16 h) at 37 °C and digested peptides were enriched with streptavidin magnetic beads (Cytiva) for 1 h at room temperature. Beads were then washed three times with DPBS and twice with HPLC grade water (Sigma Aldrich). Peptides were eluted with 50% acetonitrile (Sigma Aldrich), 0.1% formic acid (Thermo Fisher Scientific), and dried using a Speedvac (Thermo Fisher Scientific).

Cysteine-enriched peptides were reconstituted with 30% acetonitrile, 70% 200 mM EPPS pH 8.4, and labeled with 25 µg of TMT reagent (Thermo Fisher Scientific) per channel for 75 min at room temperature with rotation. Labeling was terminated by the addition of 5% hydroxylamine (Acros Organics) for 15 min followed by the addition of 10% formic acid. Peptides from four different cancer cell lines derived from the same lineage, treated with vehicle or three scout fragments, were labeled with 16 TMT tags, pooled, and dried using a Speedvac (Thermo Fisher Scientific). Peptides were then desalted with stage tips using the following procedure: peptides were reconstituted with 5% acetonitrile/0.1% formic acid and loaded onto C18 Micro Spin columns (Nest Group) pre-equilibrated with LC/MS-grade methanol (Fisher Chemical) and LC/MS-grade water containing 0.1% formic acid. C18 spin columns were washed 10 times with LC/MS grade water containing 0.1% formic acid and subsequently eluted with 80% acetonitrile, 0.1% formic acid, and dried using a Speedvac (Thermo Fisher Scientific).

MS Data Acquisition and Processing.

All mass spectrometry samples were analyzed as previously described using an Orbitrap Eclipse Tribrid Mass Spectrometer coupled with an Easy NanoLC-1200 system (Thermo Fisher Scientific) (58). Peptides were separated on a 75 µm capillary column packed with 50 cm of C18 resin (2 µm, 100 Å; Thermo Fisher Scientific) using a 180-min gradient of 4 to 35% acetonitrile in 0.1% formic acid, with a flow rate of 300 nL/min. The eluted peptides were acquired by data-dependent acquisition and quantified using the synchronous precursor selection (DDA-SPS-MS3) method for TMT quantification. MS1 spectra were acquired in the scan range of 400 to 1,400 m/z at an Orbitrap resolution of 120,000 with a maximum injection time of 50 ms and high-field asymmetric-waveform ion-mobility spectrometry (FAIMS) values set at −40, −50, and −70 compensation voltage (CV). MS2 spectra were acquired by selecting the top 20 most abundant features via collision-induced dissociation in the ion trap with an automatic gain control (AGC) setting of 10 K, a quadrupole isolation width of 0.7 m/z, and a maximum ion accumulation time of 50 ms. These spectra were relayed in real-time to an external computer for online database searching using Comet real-time searching (RTS) with a database that included mouse protein databases (release_20220301) (59, 60). Both the real-time search and the final search utilized the same forward and reverse-sequence mouse protein databases (Uniprot). Peptides were filtered using simple initial parameters: they did not match a reverse-sequence, contained TMTPro16 isobaric tags, had a maximum PPM error <50, a minimum PPM error >5, and a minimum ΔCorr of 0.10. If peptide spectra met these criteria, an SPS–MS3 scan was performed using up to 20 beta- and gamma-type fragment ions as precursors with an AGC of 250 K for a maximum of 250 ms, and a normalized HCD collision energy setting of 55 (TMTPro16).

Peptide searches were performed in Proteome Discoverer (2.5, Thermo Fisher Scientific) against the UniProt mouse protein database (release_20220301) with the following parameters: up to two missed cleavages, 20 ppm precursor tolerance, 0.6 Da fragment ion tolerance, and fully tryptic peptides were allowed with a minimum length of 6. Static modification of TMTPro16 on lysine and peptide N-termini (+304.2071 Da) and carbamidomethylation of cysteine residues (+57.0214 Da), along with oxidation of methionine (+15.9949 Da) and DBIA on cysteine residues (for iso-TMT samples) (+239.1634) as variable modifications were permitted. Results were filtered to a peptide false discovery rate (FDR) of 1%. For the TMT reporter ion quantification, all identified peptide spectral matches from the MS3 scans were extracted by an in-house program, and the reporter ion intensities were adjusted for impurity correction according to the manufacturer’s specifications. For quantification of each MS3 spectrum, a total sum signal-to-noise of all reporter ions of 100 (TMT16-plex) was used. Engagement scores were calculated as previously described (58). Raw data are available on MassIVE (MSV000098633) (61).

The supplemental sheet (Dataset S2) includes average engagement scores for each genotype (n = 3), and the results of a two-tailed unpaired Student’s t tests for engagement significance (no adjustment for multiple hypothesis testing).

Supplementation of Drinking Water with Cysteamine and Cystine.

200× cysteamine (Sigma Aldrich, M6500) and cysteine (Sigma Aldrich, 778451) solutions were made fresh in 60 mg/mL sucrose and adjusted to pH 6.0 with sodium hydroxide before being diluted to 1× in vivarium-provided acidified water, pH 6.0. Water was replaced every other day.

Cages were checked for neonates every morning. Neonates were euthanized according to approved Institutional Animal Care and Use guidelines at the Massachusetts General Hospital, weighed, and fixed for 48 h in a 10% formalin solution at room temperature. Pre-fixation, a 1 to 2 mm tail biopsy was taken for genotyping. Skin samples were taken from euthanized neonatal mice and embedded in paraffin using standard procedures. 5 μm sections were prepared from tissue blocks and stained with Hematoxylin Counterstain (Vector laboratories, H-3401) and Alcoholic Eosin Y Solution (Sigma, HT110132) for histologic analysis. Stained images were captured using a Hamamatsu Nanozoomer and analyzed using NDP.view2 software (Hamamatsu).

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Data, Materials, and Software Availability

RNAseq fastq and raw proteomic data have been deposited in GEOand MassIVE (GSE300056 (57) and MSV000098633 (61)). All other data are included in the manuscript and/or supporting information.