In vivo activities of heparan sulfate differentially modified by NDSTs during development

Abstract

Heparan sulfate proteoglycans (HSPGs) serve as co‐receptors for growth factor signaling during development. It is well known that the level and patterns of sulfate groups of heparan sulfate (HS) chains, or HS fine structures, have a major impact on HSPG function. On the other hand, the physiological significance of other structural features of HS, including NS/NA domain organization, remains to be elucidated. A blueprint of the HS domain structures is mainly controlled by HS N‐deacetylase/N‐sulfotransferases (NDSTs). To analyze in vivo activities of differentially modified HS, we established two knock‐in (KI) Drosophila strains with the insertion of mouse Ndst1 (mNdst1) or Ndst2 (mNdst2) in the locus of sulfateless (sfl), the only Drosophila NDST. In these KI lines, mNDSTs are expressed from the sfl locus, in the level and patterns identical to the endogenous sfl gene. Thus, phenotypes of Ndst1 KI and Ndst2KI animals reflect the ability of HS structures made by these enzymes to rescue sfl mutation. Remarkably, we found that mNdst1 completely rescued the loss of sfl. mNdst2 showed a limited rescue ability, despite a higher level of HS sulfation compared to HS in mNdst1 KI. Our study suggests that independent of sulfation levels, additional HS structural features controlled by NDSTs play key roles during tissue patterning.

Abbreviations

• cDNA
• complementary DNA
• CNS
• central nervous system
• dlp
• Dally-like protein
• Dpp
• Decapentaplegic
• EGF
• epidermal growth factor
• FGF
• fibroblast growth factor
• GlcNAc
• N-Acetylglucosamine
• Hh
• Hedgehog
• HS
• heparan sulfate
• HSPGs
• heparan sulfate proteoglycans
• KI
• knock-in
• KO
• knockout
• L3
• longitudinal wing vein 3
• L4
• longitudinal wing vein 4
• L5
• longitudinal wing vein 5
• mNDST1
• mouse N-deacetylase/N-sulfotransferase 1
• mNDST2
• mouse N-deacetylase/N-sulfotransferase 2
• mRNA
• messenger RNA
• NA
• N-acetylated
• NDSTs
• N-deacetylase/N-sulfotransferases
• NS
• N-sulfated
• sfl
• Sulfateless
• Sulfs
• 6-O-sulfatases
• RNAi
• RNA interference
• RT-PCR
• reverse transcription-polymerase chain reaction
• RT-qPCR
• reverse transcription quantitative polymerase chain reaction
• sgRNA
• single-guide RNA
• T2A
• Thosea asigna virus 2A
• Upd
• Unpaired
• UTR
• untranslated region
• Wg
• Wingless

INTRODUCTION

Heparan sulfate proteoglycans (HSPGs) play critical roles during development and pathogenesis. Importantly, they function as co-receptors for growth factor signaling, regulating distribution and reception of secreted signaling proteins.^1–5 In Drosophila, HSPGs regulate gradient formation and signaling of four key morphogen molecules, Decapentaplegic (Dpp; a Drosophila BMP), Wingless (Wg; a Drosophila Wnt), Hedgehog (Hh), and Unpaired (Upd; a ligand of the Jak/Stat pathway), as well as other secreted factors, such as FGFs, Vein (a EGF receptor ligand), and Slit.⁶

The biological function of HSPGs is dependent on both core protein and sugar moieties. During HS biosynthesis, HS undergoes sequential modification events. The first step, N-deacetylation and N-sulfation of GlcNAc units, is catalyzed by HS N-deacetylase/N-sulfotransferase (NDST). Human and mouse have four NDSTs while Drosophila has a single NDST gene, sulfateless (sfl).⁷ Since this reaction is essential for subsequent HS modifications, loss of sfl eliminates most, if not all, HS activity.^7,8 In the following steps, C5-epimerase converts glucuronic acid residues to iduronic acid and O-sulfotransferases add sulfate groups to the growing chain. After the HS modification steps in the Golgi, HS can be further modified extracellularly by a family of extracellular HS 6-O-endosulfatases (Sulfs).⁹ Sulfs remove a subset of 6-O sulfate groups within the highly sulfated domains of HS.¹⁰ Drosophila Sulf1 negatively regulates Wg, Hh, and EGFR signaling by reducing the number of ligand binding sites on HS.^11–14 All these pathways are inhibited by Sulf1 overexpression and upregulated in Sulf1 mutants. Therefore, a higher level of HS sulfation is believed to generally upregulate growth factor signaling in Drosophila. Since only a fraction of potential target units are modified in each biosynthetic step, the resulting HS chains have remarkable levels of structural heterogeneity. The HS fine structures thus generated have a major impact on HSPG function.¹⁵

In addition to the fine structures, HS has other structural organizations. For example, chain length, net charge, and the degree/distribution of C5-epimerization, each of which varies depending on tissues and species. Furthermore, HS has regions that are highly sulfated (NS domains) interspersed with nonsulfated regions (NA domains).^16,17 However, the biological importance of these structural features is poorly understood.

The formation of the NS/NA-domain structures is mainly controlled by NDSTs. When an NDST adds sulfate groups to a stretch of the growing chain, other enzymes modify that region, leading to the formation of an NS domain. When an NDST does not add N-sulfate groups, it will generate an NA domain. In most human and mouse tissues, Ndst1 plays a major role in the production of HS, composed of discreet NS and NA domains.^18,19 Ndst1 knockout (KO) mice show a variety of severe defects in the development of many tissues, including lungs, eyes, skeleton, and vasculatures, resulting in high levels of lethality.^20–22 In tissue culture cells, another NDST enzyme, named Ndst2, shows a higher activity of N-sulfation than Ndst1, producing a higher degree of N-sulfation and longer stretches of NS domains.^23,24 This is consistent with its in vivo function. Ndst2 is responsible for the production of heparin, which lacks the alternating patterns of these domains and is characterized as a continuous NS domain.^17,25,26 The only defect of Ndst2 KO mice is the lack of heparin in mast cells, leading to abnormal mast cell function.²⁵ Thus, in vitro and in vivo studies of Ndst1 and Ndst2 revealed differential enzymatic activities and distinct developmental roles of these enzymes. However, as these NDSTs regulate structures of two different classes of polysaccharides, it is not feasible to directly compare their in vivo activities.

Here, we devised a heterologous system to measure the in vivo activities of the different HS structures generated by Ndst1 and Ndst2. Mouse Ndst1 (mNdst1) and mouse Ndst1 (mNdst2) knock-in (KI) strains in the sfl locus allow us to compare the ability of the two mouse NDSTs to rescue the loss of sfl. We found that mNdst1 shows a remarkably higher level of in vivo activity in Drosophila compared to mNdst2 in spite of a lower sulfation level of HS. Our observations suggest that HS sulfation pattern design controlled by NDSTs, not simply sulfation levels, plays important roles in vivo.

MATERIALS AND METHODS

Drosophila strains

The following fly strains were used in this study: Oregon-R and UAS-sfl RNAi (HMS00543, BDSC #34601). Flies were raised on a standard cornmeal fly medium at 25°C, unless otherwise indicated.

Generation of sfl^KI:mNdst1 and sfl^KI:mNdst2

cDNA clones for mNdst1 and mNdst2 were obtained from Lena Kjellén (Uppsala University). To generate flies expressing mNdst1 or mNdst2 instead of sfl, we inserted the sequence containing T2A followed by mNdst1 or mNdst2 CDS and sfl 3′-UTR at the sfl locus using CRISPR/Cas9-mediated homology-directed repair.^27,28 The single-guide RNA (sgRNA) was constructed by annealing 5′-CTTCGCTGTTGGACAAATACTGCC-3′ and 5′-AAACGGCAGTATTTGTCCAACAGC-3′ and ligating in the BbsI-digested pU6-BbsI-chiRNA plasmid (a gift from Melissa Harrison, University of Wisconsin; Kate O'Connor-Giles, Brown University; and Jill Wildonger, University of Wisconsin; Addgene #45946).^27,28 To generate the repair template, sfl homology arms, T2A, mNdst1 (or mNdst2) CDS, and sfl 3′-UTR were assembled using NEBuilder HiFi DNA Assembly Master Mix (E2611S; New England Biolabs). A mixture of 50 ng/µL of the sgRNA plasmid and 250 ng/µL of the repair template was injected into the vasa-Cas9 embryos (BDSC #51323) by GenetiVision. The homologous recombinants were screened by PCR and verified by Sanger sequencing. The obtained strains were backcrossed to Oregon-R for five generations.

Preparation of adult wings and legs

The right wings from female flies were dehydrated in ethanol and subsequently with xylene.^29,30 Adult legs were boiled in 2.5 N sodium hydroxide, washed in distilled water, and dehydrated in 2-propanol.²⁹ The specimens were mounted in Canada balsam (Benz Microscope, BB0020).

RT-PCR

Expression of sfl, mNdst1, and mNdst2 was analyzed by RT-PCR. Actin5C was used as a control. Wild-type, Nsdt1, and Ndst2 adult flies were homogenized in 300 μL of TRIzol® reagent (Invitrogen; 15596-026), and total RNA was isolated using Direct-zol™ RNA MiniPrep (Zymo Research; R2050). cDNA was synthesized from 50 ng of total RNA using SuperScript® III First-Strand Synthesis System for RT-PCR (Invitrogen; 18080-051). A 0.5 μL aliquot of the cDNA synthesis reaction mixture was used to amplify the target cDNAs using the following PCR primers:

sfl (forward): 5′-CAAACGAAGTCATGCCCTGC-3′.

sfl (reverse): 5′-CAGATCCTTCAGTGCCCTCG-3′.

mNdst1 (forward): 5′-TGGATTCCCGAGCCTTCCTA-3′.

mNdst1 (reverse): 5′-ACCTGGTGTTCTGGAGGTCT-3′.

mNdst2 (forward): 5′-TCCCTGTTCCTTCCAATGCC-3′.

mNdst2 (reverse): 5′-TACCAGGAGTAGGCCCTGTC-3′.

Act5C (forward): 5′-GGCGCAGAGCAAGCGTGGTA-3′.

Act5C (reverse): 5′-GGGTGCCACACGCAGCTCAT-3′.

PCR products with expected size were analyzed by agarose gel electrophoresis.

For RT-qPCR, RNA samples were prepared from the whole third-instar larvae. cDNA was synthesized using SuperScript III First-Strand (Invitrogen). qPCR assays were performed for three independent biological replicates in a LightCycler 480 Instrument II (Roche) using LightCycler 480 SYBR Green I Master (Roche). Expression of Ribosomal protein L23 (RpL23rpL23) was used for normalization. Fold changes were calculated using the ΔΔCt method.

In situ RNA hybridization

In situ RNA hybridization was performed as described previously.³¹ Wing imaginal discs were dissected from third-instar larvae and fixed with 4% paraformaldehyde. Digoxigenin-labeled RNA probes were synthesized using a DIG RNA Labeling kit (Roche Applied Science). The hybridized probes were detected by anti-digoxigenin antibody conjugated with alkaline phosphatase (Roche Applied Science). The signal was developed by a standard protocol using 3,3-diaminobenzidine as a substrate.

Disaccharide analysis

Approximately 10 mg of lyophilized larvae were used for analysis of HS. Briefly, crude glycosaminoglycans were obtained by extraction with 0.5% sodium dodecyl sulfate, 0.1 M NaOH, and 0.8% NaBH₄, followed by ethanol precipitation. Chondroitin sulfate was removed from the crude glycosaminoglycan solution by chondroitinase treatment, followed by separation with a centrifugal ultrafiltration membrane, NANOSEP 3K OMEGA (Pall Life Science). The HS sample was digested with a heparitinase mixture (Seikagaku), and the resulting disaccharide species were separated using reversed-phase ion-pair chromatography. The effluent was monitored fluorometrically for postcolumn detection of HS disaccharides.⁸

RESULTS

mNdst KI alleles in the sfl locus

Drosophila produces HS, but not heparin, using a single NDST homolog, Sfl. In sfl mutants, all HS-dependent growth factor pathways are disrupted, and zygotic null mutants are fully lethal by late larval stages. Based on sequence homology analyses, the similarities of the Sfl amino acid sequence to those of mNdst1 and mNdst2 are comparative (Supporting Information S1: Figure S1). For example, sequence alignment using EMBOSS Water with the Smith–Waterman algorithm revealed 70.2% sequence similarity between Sfl-Ndst1 and 71.2% between Sfl-Ndst2. Thus, structurally, mNdst1 and mNdst2 are equally similar to (and distant from) Sfl.

We devised a system to compare the in vivo activities of HS modified by different NDSTs using the Drosophila model. We generated two KI Drosophila strains with the insertion of mNdst1 or mNdst2 in the endogenous sfl locus via CRISPR/Cas9 gene editing (Figure 1A). To achieve expression of mNdst1 and mNdst2 in the patterns of the endogenous sfl gene, we employed the Thosea asigna virus 2A (T2A) in-frame fusion technology, which utilizes a “ribosomal skipping” mechanism of the T2A peptide.^32,33 We inserted T2A–mNdst complementary DNA (cDNA) constructs after Tyr₃₁₅ of the sfl protein coding sequence in frame, and thus mNdst cDNAs are transcribed as a fusion RNA. T2A is a viral ribosomal skipping site that terminates translation at the beginning of the peptide and reinitiates it after the site, producing a truncated Sfl peptide and full-length mNdst enzymes. mNdst protein coding sequences are followed by the 3′-untranslated region (3′-UTR) of sfl messenger RNA (mRNA), including the polyadenylation sequence. Therefore, the sfl genomic sequence after the insertion site is not transcribed. As a result, no functional Sfl protein is produced in these transgenic strains (Supporting Information S1: Figure S1). Instead, the mNdsts are expressed from the sfl locus in the level and patterns identical to the endogenous sfl gene in wild type. Reverse transcription-polymerase chain reaction (RT-PCR) analyses verified the loss of sfl mRNAs and expression of respective Ndst genes in these KI alleles (Figure 1B). In situ RNA hybridization showed that sfl mRNA is uniformly expressed in the wild-type larval imaginal discs and central nervous system (CNS), consistent with the fact that sfl is an essential gene required for the biosynthesis of ubiquitous HS. We detected transcripts of Ndst1 in Ndst1 KI and Ndst2 in Ndst2 KI at indistinguishable levels and patterns from sfl (Supporting Information S1: Figure S2A–H). RT-quantitativePCR (RT-qPCR) also confirmed comparative levels of expression of sfl in wild type, Ndst1 in Ndst1 KI, and Ndst2 in Ndst2 KI (Supporting Information S1: Figure S2I). These novel Drosophila strains, sfl^KI:mNdst1 and sfl^KI:mNdst2, are referred to as Ndst1 and Ndst2 strains, respectively, in this paper. This heterologous system allows the first systematic approach to compare in vivo activities of distinct HS structures modified by different NDSTs.

[IMAGE OMITTED. SEE PDF]

Disaccharide analyses of HS from Ndst KI alleles

As detailed below, we found that mNdst1 and mNdst2 rescue the loss of sfl in these KI alleles, and animals survive to adult stages. To determine if Ndst1 and 2 differentially modify HS in Drosophila, as has been observed in vitro,²⁴ we analyzed the structures of “Drosophila HS” modified by mNdst1 and 2 by disaccharide analysis. Briefly, HS was purified from third-instar larvae and completely digested into disaccharides by heparitinase. The resultant disaccharide species were separated and quantified by reversed-phase ion-pair chromatography with a postcolumn detection system.^8,11,34,35

As was previously shown,³⁶ HS from sfl mutant larvae lacks sulfated disaccharide units (Table 1 and Figure 2A). mNdst1 and 2 partially rescued this phenotype, producing sulfated disaccharide species, including the tri-sulfated unit (ΔUA2S-GlcNS6S or 2SNS6S). However, the sulfation levels of HS in these KI alleles were significantly lower than in wild type. The proportion of sulfated disaccharides of Ndst1 and Ndst2 were 23% and 31% of wild type, respectively (Table 1 and Figure 2B).

Table 1 Disaccharide analyses of HS from wild-type, sfl, Ndst1, and Ndst2 animals.

Genotype	HS (unsaturated disaccharide, %)	Total HS (ng/mg dry tissue)
NAc	NS	NAc6S	NS6S	2SNS	2SNS6S
Wild type	43.8 ± 1.7	18.6 ± 0.9	1.6 ± 0.4	21.2 ± 0.4	10.2 ± 0.4	4.6 ± 0.1	18.1 ± 1.2
sfl	100.0	ND	ND	ND	ND	ND	13.6 ± 0.7
Ndst1	86.0 ± 2.5	6.5 ± 1.3	1.2 ± 0.2	3.5 ± 0.4	1.4 ± 0.5	1.4 ± 0.2	17.7 ± 1.3
Ndst2	81.4 ± 1.8	8.5 ± 0.8	1.7 ± 0.1	5.1 ± 0.7	1.7 ± 0.4	1.7 ± 0.1	15.2 ± 1.6

[IMAGE OMITTED. SEE PDF]

Importantly, consistent with previous in vitro observations,^23,24 Ndst2 HS has significantly higher levels of sulfated disaccharides compared to that from Ndst1 (Table 1 and Figure 2B). The average percentages are higher in Ndst2 than Ndst1 for all sulfated disaccharides. The percentage of total sulfated disaccharides was significantly higher in Ndst2 than Ndst1 (Table 1). We also calculated the relative amounts of total sulfate groups in each genotype by multiplying the ratios of monosulfated (NS, 6S), disulfated (NS6S, 2SNS), and trisulfated (2SNS6S) disaccharides by 1, 2, and 3, respectively (Figure 2C). This value was also significantly higher in Ndst2, indicating that Ndst2 HS contains more sulfate groups compared to Ndst1. Our in vivo results confirmed that these two enzymes differentially modify HS in Drosophila. More specifically, mNdst2 shows a higher activity to add sulfate groups on HS compared to mNdst1 in this heterologous system.

Comparison of total HS (ng/mg dry tissue) recovered in disaccharide analyses showed no significant difference in the amount of HS from wild-type and the two Ndst KI alleles (Table 1 and Figure S3). Interestingly, the amount of unsulfated polysaccharide (heparosan) from sfl mutants was lower than HS from other genotypes. As an NDST affects HS chain length,³⁷ this may reflect a shorter chain length of sfl heparosan. Future studies are needed to reveal the molecular basis for the reduced recovery of heparosan.

mNdst1 fully rescues the lethality of sfl

To examine in vivo activities of differentially modified HS by Ndst1 and Ndst2, we asked which gene shows a higher ability to rescue the lethality of sfl mutant. Our lethality assay showed that both genes can rescue it but to significantly different degrees. We found that the lethality of sfl mutants is completely restored by Ndst1 (Figure 3A). This is remarkable given that the HS sulfation level of these animals is only 23% of wild type. Thus, the minimal level of HS sulfation required for normal viability is unexpectedly low.

[IMAGE OMITTED. SEE PDF]

In contrast, Ndst2 KI shows a high level of lethality (91.4% for females, 74.6% for males) (Figure 3A), indicating a limited ability of the rescue. We found that the lethality was also fully rescued in Ndst1/Ndst2 heterozygotes (Figure 3A), which suggests a dominant effect of Ndst1's ability to complement sfl function.

Ndst2 KI shows leg patterning defects

Ndst2 homozygotes die at various stages, including the lethality at late pupal/pharate adult stages due to eclosion failure. We found that dead pharate adults dissected from pupal cases have malformed legs. In addition, Ndst2 adult flies that successfully eclosed also show morphological defects in their legs. Ndst2 adults exhibit patterning abnormalities in fore, middle, or hindlegs, such as twisting and bending of the appendages (Figure 3B–G). In contrast, these defects are not observed in Ndst1. The leg patterning phenotypes explain the eclosion failure and appear to contribute to the lethality of Ndst2 homozygotes to some extent. Except for the legs, Ndst2 survivors do not show any obvious major defect in other adult organs, suggesting that leg morphogenesis is particularly sensitive to a change in N-sulfation patterns of HS. At this point, however, the molecular basis for this sensitivity of leg morphogenesis to Ndst2 KI is unknown.

Wing morphology of Ndst1 KI and Ndst2 KI adult survivors

To further examine the effect of HS sulfation design mediated by Ndsts, we compared in vivo patterning activities of Ndst1 and Ndst2 during wing development, a model system extensively used for morphogen function studies. When sfl function is compromised by RNA interference (RNAi) knockdown specifically in the developing wing using Bx-Gal4, a commonly used wing Gal4 driver, knockdown animals (Bx>sfl RNAi) show a variety of wing phenotypes, including loss of anterior and posterior crossveins, wing cell growth defects, wing notching, ectopic venation, and blistered wing (Figure 4B,C). These patterning defects are consistent with the idea that sfl is required for Hh, Dpp, and Wg morphogen signaling during wing development.^7,38–40 We found that overall wing patterning of adult survivors of Ndst1 and Ndst2 is quite normal (Figure 4D,E), indicating a remarkable ability of both genes to rescue the loss of sfl in the wing. It is particularly interesting that Ndst2 wings do not show any obvious morphological defects (Figure 4E) despite its high level of lethality.

[IMAGE OMITTED. SEE PDF]

However, quantitative analyses of wing morphology parameters of Ndst1 and Ndst2 adults revealed that in general, Ndst1 shows a stronger activity to rescue sfl patterning defects compared to Ndst2 (Figure 4F–I). The following parameters were used for comparison: wing area, length, width, and aspect ratio (width/length). Wing length was measured as the distance between the distal edge of the third longitudinal vein (L3) to the wing hinge (horizontal arrow in Figure 4A). Wing width was measured as the shortest distance between the posterior edge of the fifth longitudinal vein (L5) to the wing margin (vertical arrow in Figure 4A).

We found that both Ndst1 and Ndst2 wings are smaller, with reduced length and width (Figure 4F). Ndst2 is more severely affected than Ndst1 in the wing area, length, and width (Figure 4F–H). Interestingly, however, Ndst2 wings are particularly short in length compared to both wild type and Ndst1 (Figure 4G). As a result, the wing aspect ratio (width/length) of Ndst2 wings is significantly higher than wild type and Ndst1, which means that Ndst2 wings are rounder (Figure 4I). On the other hand, Ndst1 wings have a wild-type wing aspect ratio, indicating that length and width are proportionally reduced.

Taken together, Ndst KI alleles show not only different levels of rescue of sfl mutation but also distinct patterns of complementation. During wing development, some morphogen pathways regulate growth and patterning along the A–P axis (Hh and Dpp), and others affect the D–V axis (Wg). Therefore, our data suggest a possibility that HS designs differentially modified by NDSTs may uniquely affect distinct morphogen pathways.

Wing margin phenotypes of Ndst1 KI and Ndst2 KI adult flies

We next determined if mNdst1 and mNdst2 show different activities on the wing margin formation, an event controlled by Wg signaling. The anterior wing margin bears several rows of different types of sensory organs: thick and short mechanosensory bristles at the edge and thinner chemosensory bristles slightly posterior to the edge (Figure 5A⁴¹). Dally, one of the two Drosophila glypicans, acts as a Wg co-receptor, and the loss of dally substantially decreases the number of these bristles. On the other hand, knockdown of dally-like protein (dlp), the second glypican gene, results in increased Wg signaling near the wing margin.^42–46 This leads to the formation of ectopic mechanosensory bristles that are shifted from their normal position, the edge of the wing.⁴²

[IMAGE OMITTED. SEE PDF]

We found that wings of both Ndst1 and Ndst2 adult survivors have a normal number of chemosensory bristles at the anterior wing margin (Figure 5A–D). Interestingly, however, Ndst2 KI wings bear ectopic mechanosensory bristles, similar to dlp knockdown (Figure 5C,E). Ndst1 occasionally showed this phenotype, but the difference between wild-type and Ndst1 wings was not statistically significant (Figure 5B,E). These observations suggest that HS modified by mNdst2 may induce an abnormally higher level of Wg signaling at the wing margin.

Different effects of Ndst1 and Ndst2 on the ratio of the L3/L4 domain

Although Ndst1 shows much higher activities to rescue the loss of sfl in most contexts, there is a phenotype in which Ndst2 may function better than Ndst1. Hh signaling regulates the growth and patterning of the central region of the wing, and manipulation of this pathway affects the spacing between longitudinal wing veins L3 and L4 (Figure 4A). Thus, the ratio of the area between L3 and L4 (L3/L4 domain) in the total wing area reflects Hh signaling activity.⁴⁷

We found that the L3/L4 domain ratio is significantly reduced in Ndst1 wings, suggesting that Ndst1 shows a low activity to rescue reduced Hh signaling of the sfl mutants (Figure 5F). In contrast, in Ndst2 wings, the L3/L4 ratio was not significantly different from wild type. Thus, unlike other parameters, Ndst2 appears to show a higher ability to rescue an Hh-dependent process compared to Ndst1.

DISCUSSION

The first step of HS modification is catalyzed by NDSTs, which generate a blueprint of domain structures of HS. In mammals, HS has alternating patterns of a highly sulfated NS domain and an unsulfated NA domain. Different NDSTs generate distinct NS/NA domain structures. For example, Ndst1 (which is critical for producing HS) makes clearly distinct NS/NA domains, while Ndst2 (which produces heparin) continuously adds sulfate groups to make a chain composed of a single NS domain, without NA domains.

It is worth noting that Ndst2 can synthesize HS-like structures in vivo.^48,49 In wild-type mouse embryos, Ndst2 is present in the liver but it does not contribute to HS structure. However, in Ndst1 KO liver cells, where Ndst2 is the only NDST isoform expressed, it forms the same HS structures as wild type. The observation suggested that the GAGosome, an active complex of the HS biosynthetic/modifying enzymes, preferentially incorporates Ndst1, but Ndst2 is recruited to the GAGosome in the absence of Ndst1.

In vivo functions of Ndst1 or Ndst2 have been extensively studied using KO mouse strains.^{18,20–22,25} Ndst1 KO mice are perinatal lethal and show developmental defects in the respiratory, nervous, skeletal, and vascular systems.²² The broad array of phenotypes from Ndst1 KO mice include—but are not limited to—abnormal lungs resembling respiratory distress syndrome in humans, cerebral hypoplasia and patterning defects, axon guidance defects, neural tube closure failure, the loss of neural crest cell-derived elements, the loss of olfactory bulbs, delayed ossification, defects of the mandible, and delayed pericyte recruitment.^18,21,50,51 Conversely, Ndst2 KO mice are viable and only show one clear phenotype: disrupted mast cells due to the loss of sulfated heparin synthesized by NDST2.^18,52 The lack of lethality and phenotypes observed in Ndst2 KO mice is striking given that Ndst2, along with Ndst1, is expressed throughout the body in mice.¹⁸ The difference between Ndst1 KO and Ndst2 KO mice provides clear evidence that Ndst1 and Ndst2 function distinctly in heparan sulfate biosynthesis.

However, since Ndst1 or Ndst2 contribute to completely different biological roles, the mouse KOs cannot allow us to directly compare the activities of HS differentially modified by these enzymes. To fill this knowledge gap, we generated two KI strains with an insertion of mNdst1 or mNdst2 in the endogenous sfl locus using the T2A in-frame fusion technology. This methodology has been successfully used to generate KI transgenic strains in Drosophila, including with high-throughput genome-wide platforms.^53,54 Our heterologous system provides a unique opportunity to study the in vivo activities of HS differentially modified by mNDST1 and mNDST2. Interestingly, the ability of these mouse genes to rescue sfl mutants was significantly different. Ndst1 almost perfectly rescues the lethality and major morphological defects of sfl mutants. In contrast, Ndst2 shows limited ability to rescue them. Our findings were remarkable for two reasons. First, HS in Ndst1 bears only 23% of the wild-type sulfation level (Table 1 and Figure 2), but exhibits virtually wild-type in vivo activities during tissue patterning. Second, although Ndst2 KI HS has higher compositions of sulfated disaccharide species compared to Ndst1, consistent with their enzymatic activities, it shows a significantly lower activity to rescue sfl. Our observations suggest a possibility that HS structures controlled by NDSTs, rather than simply sulfation levels, are important for in vivo patterning activities of HS.

Although Ndst1 generally showed a significantly higher ability to rescue sfl compared to Ndst2, there were a few exceptions. First, the wing aspect ratio (width/length) of Ndst1 and Ndst2 are different, making Ndst2 wings rounder. Second, the L3/L4 domain ratio of Ndst1 wings is more significantly reduced. These observations suggest that HS structures modified by Ndst1 and 2 may have unique and differential effects on distinct morphogen pathways.

We still do not know whether the NS/NA domain structure is the source of differential in vivo activities between Ndst1 and Ndst2. What specific HS structural features are differentially modified by these enzymes in Drosophila remains to be determined. It had been difficult to analyze Drosophila HS structures, with the exception of disaccharide composition, mainly due to the difficulty of metabolic radiolabeling of HS in vivo using Drosophila animals.^{11–13,31,34,55,56} We recently established a novel system to overcome this obstacle.³⁵ This platform is composed of two steps: generation of immortalized cell lines derived from existing Drosophila strains^57–59 and metabolic radiolabeling and structural analysis of HS from these established cell lines.³⁵ This system allows us to connect in vivo phenotypic information of Drosophila strains and detailed HS structural features. Future studies using this strategy will be able to identify the differences in HS structures between Ndst1 and Ndst2 animals. Thus, our study opens up a new strategy of “in vivo structure–function analysis” of HS.

AUTHOR CONTRIBUTIONS

Conceptualization: Eriko Nakato, Masahiko Takemura, and Hiroshi Nakato. Investigation: Eriko Nakato, Sarah Baker, Akiko Kinoshita-Toyoda, Collin Knudsen, Yi-Si Lu, Masahiko Takemura, Hidenao Toyoda, and Hiroshi Nakato. Writing: Eriko Nakato, Hidenao Toyoda, and Hiroshi Nakato. Review and editing: Collin Knudsen and Hiroshi Nakato. Funding acquisition: Hiroshi Nakato. Supervision: Hiroshi Nakato and Hidenao Toyoda.

ACKNOWLEDGMENTS

We thank Lena Kjellén, Melissa Harrison, Kate O'Connor-Giles, Jill Wildonger, and the Bloomington Drosophila Stock Center for cDNA clones, plasmid vectors, and fly strains. We are grateful to Elaina Creagan, Kara Eckberg, Kristin Grandt, and Tomomi Izumikawa for their support of Drosophila genetic experiments and sample preparations. This work was supported by the National Institutes of Health (R35 GM131688 to H. N.).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT

All data are contained within the manuscript.

ETHICS STATEMENT

The authors confirm that the ethical policies of the journal as noted on the journal's author guidelines have been adhered to.

References

U. Lindahl, J.‐p Li, Int. Rev. Cell. Mol. Biol. 2009, 276, 105.

C. A. Kirkpatrick, S. B. Selleck, J. Cell Sci. 2007, 120, 1829.

J. D. Esko, S. B. Selleck, Annu. Rev. Biochem. 2002, 71, 435.

J. P. Li, M. Kusche‐Gullberg, Int. Rev. Cell. Mol. Biol. 2016, 325, 215.

D. Xu, J. D. Esko, Annu. Rev. Biochem. 2014, 83, 129.

H. Nakato, J. P. Li, Int. Rev. Cell. Mol. Biol. 2016, 325, 275.

X. Lin, N. Perrimon, Nature 1999, 400, 281.

H. Toyoda, A. Kinoshita‐Toyoda, S. B. Selleck, J. Biol. Chem. 2000, 275, 2269.

G. K. Dhoot, M. K. Gustafsson, X. Ai, W. Sun, D. M. Standiford, C. P. Emerson, Science 2001, 293, 1663.

M. Morimoto‐Tomita, K. Uchimura, Z. Werb, S. Hemmerich, S. D. Rosen, J. Biol. Chem. 2002, 277, [eLocator: 49175].

A. Kleinschmit, T. Koyama, K. Dejima, Y. Hayashi, K. Kamimura, H. Nakato, Dev. Biol. 2010, 345, 204.

A. Wojcinski, H. Nakato, C. Soula, B. Glise, Dev. Biol. 2011, 358, 168.

A. Kleinschmit, M. Takemura, K. Dejima, P. Y. Choi, H. Nakato, J. Biol. Chem. 2013, 288, 5081.

J. P. Butchar, D. Cain, S. N. Manivannan, A. D. McCue, L. Bonanno, S. Halula, S. Truesdell, C. L. Austin, T. L. Jacobsen, A. Simcox, Genetics 2012, 191, 1213.

H. Nakato, K. Kimata, Biochim. Biophys. Acta 2002, 1573, 312.

J. D. Esko, U. Lindahl, J. Clin. Invest. 2001, 108, 169.

P. Carlsson, J. Presto, D. Spillmann, U. Lindahl, L. Kjellén, J. Biol. Chem. 2008, 283, [eLocator: 20008].

E. Forsberg, L. Kjellén, J. Clin. Invest. 2001, 108, 175.

L. Kjellén, Biochem. Soc. Trans. 2003, 31, 340.

K. Grobe, J. Ledin, M. Ringvall, K. Holmborn, E. Forsberg, J. D. Esko, L. Kjellén, Biochim. Biophys. Acta 2002, 1573, 209.

K. Grobe, M. Inatani, S. R. Pallerla, J. Castagnola, Y. Yamaguchi, J. D. Esko, Development 2005, 132, 3777.

M. Ringvall, L. Kjellen, Prog. Mol. Biol. Transl. Sci. 2010, 93, 35.

W. F. Cheung, I. Eriksson, M. Kusche‐Gullberg, U. Lindahl, L. Kjellén, Biochemistry 1996, 35, 5250.

D. S. Pikas, I. Eriksson, L. Kjellén, Biochemistry 2000, 39, 4552.

E. Forsberg, G. Pejler, M. Ringvall, C. Lunderius, B. Tomasini‐Johansson, M. Kusche‐Gullberg, I. Eriksson, J. Ledin, L. Hellman, L. Kjellén, Nature 1999, 400, 773.

P. Carlsson, L. Kjellen, Handb. Exp. Pharmacol. 2012, 60, 23.

S. J. Gratz, F. P. Ukken, C. D. Rubinstein, G. Thiede, L. K. Donohue, A. M. Cummings, K. M. O'connor‐Giles, Genetics 2014, 196, 961.

X. Ren, Z. Yang, J. Xu, J. Sun, D. Mao, Y. Hu, S.‐J. Yang, H.‐H. Qiao, X. Wang, Q. Hu, P. Deng, L.‐P. Liu, J.‐Y. Ji, J. B. Li, J.‐Q. Ni, Cell Rep. 2014, 9, 1151.

M. Fujise, S. Izumi, S. B. Selleck, H. Nakato, Dev. Biol. 2001, 235, 433.

M. Takemura, F. Noborn, J. Nilsson, N. Bowden, E. Nakato, S. Baker, T. Y. Su, G. Larson, H. Nakato, Mol. Biol. Cell 2020, 31, 813.

K. Dejima, M. Takemura, E. Nakato, J. Peterson, Y. Hayashi, A. Kinoshita‐Toyoda, H. Toyoda, H. Nakato, J. Biol. Chem. 2013, 288, [eLocator: 34384]. [DOI: https://dx.doi.org/10.1074/jbc.m113.499269]

J. A. Bosch, R. Colbeth, J. Zirin, N. Perrimon, Genetics 2020, 214, 75.

K. Kamimura, M. Fujise, F. Villa, S. Izumi, H. Habuchi, K. Kimata, H. Nakato, J. Biol. Chem. 2001, 276, [eLocator: 17014].

K. Dejima, A. Kleinschmit, M. Takemura, P. Y. Choi, A. Kinoshita‐Toyoda, H. Toyoda, H. Nakato, J. Biol. Chem. 2013, 288, 6574.

A. Simcox, Methods Mol. Biol. 2013, 945, 1.

A. Simcox, S. Mitra, S. Truesdell, L. Paul, T. Chen, J. P. Butchar, S. Justiniano, PLoS Genet. 2008, 4, [eLocator: e1000142].

A. A. Simcox, C. Austin, T. Jacobsen, H. Jafar‐Nejad, Fly 2008, 2, 306.