Characterization of the primary structure of the major silk gene, h-fibroin, across caddisfly (Trichoptera) suborders

Summary

Larvae of caddisflies (Trichoptera) produce silk to build various underwater structures allowing them to exploit a wide range of aquatic environments. The silk adheres to various substrates underwater and has high tensile strength, extensibility, and toughness and is of interest as a model for biomimetic adhesives. As a step toward understanding how the properties of underwater silk evolved in Trichoptera, we used genomic data to identify full-length sequences and characterize the primary structure of the major silk protein, h-fibroin, across the order. The h-fibroins have conserved termini and basic motif structure with high variation in repeating modules and variation in the percentage of amino acids, mainly proline. This finding might be linked to differences in mechanical properties related to the different silk usage and sets a starting point for future studies to screen and correlate amino acid motifs and other sequence features with quantifiable silk properties.

Graphical abstract

Highlights

• •New sequences of the major caddisfly silk protein, h-fibroin, are reconstructed
• •Repetitive modules vary widely across clades with different silk use
• •H-fibroin of retreat/capture net makers have higher proportion of proline
• •Amino acid differences may be linked to mechanical properties of silk

Entomology; Genetics; Molecular biology

Introduction

Insects use silk for a variety of purposes^1,2 and research has focused on terrestrial Lepidoptera (moths and butterflies), especially on the commercially important silkworm Bombyx mori.^{3,4,5,6,7,8,9,10,11} Less well studied are aquatic insects that use silk. These include the most speciose primary aquatic insect order, Trichoptera (caddisflies), which exhibit diverse silk usage strategies.¹² Similar to their sister order Lepidoptera, several caddisfly species construct cocoons for metamorphosis in the final larval instar.¹³ In earlier larval stages, they produce a diverse array of underwater structures, which is reflected in their phylogeny.^13,14 Trichoptera is divided into two suborders, which are distinguished by differences in morphology, habitat, and use of silk¹⁴: Integripalpia (cocoon and tube case makers) and Annulipalpia (fixed retreat makers). Further differentiation in silk use also occurs at the subordinal and superfamily level (reviewed in¹³). Basal Integripalpia, which show diverse case-making behaviors (free-living, tortoise case-making, purse case-making), are referred to as cocoon makers because they pupate in a silken pupal cocoon,¹⁵ whereas larvae of tube case-making Integripalpia build portable, tubular cases made purely from silk or from diverse materials encountered in their habitats, such as small stones or plant materials that are “taped” together with silk. The cases provide protection and channel oxygenated water past the body allowing them to obtain oxygen in lentic environments.¹⁵ They are of various shapes and materials and often species-specific, i.e., distinctive among congeneric species (i.e., Micrasema longulum: silk case, Micrasema setiferum: sand-stone material, Micrasema wataga: plant material). Larvae of Annulipalpia create stationary shelters of silk, often with mineral particles or plant material, which are fixed to the substrate (e.g., stones or aquatic plants). These retreats serve primarily as physical protection against predation. In addition, members of Annulipalpia construct various kinds of nets to filter suspended organic particles from flowing water or to catch small invertebrates.¹³

Caddisfly silk is a tough adhesive multi-network fiber.^16,17 Studies on the mechanical properties of single silk fibers revealed that they are viscoelastic, have an extensibility of >100%,¹⁸ display large strain cycle hysteresis, and self-recover 99% of their initial stiffness and strength.¹⁶ Caddisfly silk is of practical interest as a model for the development of biomimetic materials with applications in aqueous environments,^17,19 such as the human body (e.g., in medicine for tissue engineering, as tough hydrogels,²⁰ surgical structures or bio-bandages²¹).

Their diverse net- and case-making behavior allow caddisflies to exploit a range of ecological niches, which raises the question of how the properties of underwater silk evolved in Trichoptera. To begin to answer this question, and to gauge the potential of caddisfly silk in material sciences, a comprehensive study of the primary molecular structure of the major silk protein, h-fibroin, is necessary.

The silk fiber in Trichoptera consists of two filaments derived from a pair of labial glands (see also²²). As in most Lepidoptera, the fiber core is assembled from a large (200–500 kDa) heavy-chain fibroin (h-fibroin) and the smaller (∼25 kDa) light-chain fibroin (l-fibroin) protein.²³ In contrast to the silk of the silkworm B. mori, a homolog of glycoprotein P25 (fibrohexamerin) has not been detected in Trichoptera²³ and, instead of being surrounded by a sericin layer, the central fiber core is covered with a thinner, poorly characterized peripheral layer.²⁴ Additional silk fiber components were identified by transcriptome and proteome analysis of the silk gland (e.g., peroxinectin, a novel structural component with sequence similarity to the elastic PEVK region of the muscle protein titin, and mucins).^25,26 The h-fibroin is the major silk protein by size and mass. On a macroscale level, the structural organization of the trichopteran h-fibroin is similar to Lepidoptera.²³ It consists of non-repetitive amino(n)- and carboxyl(c)-terminal domains flanking a central region, composed of repeated structural modules.^17,27 However, the central regions exhibit no sequence conservation between orders.²⁸ In Trichoptera, these consist of repeating (SX)_nE motifs in which the S (serine) is often phosphorylated^19,29,30 and where X is primarily an amino acid with hydrophobic or aromatic side chains and sometimes arginine, E is glutamic acid, and n is 3–5. The (SX)_n E motifs are separated by glycine-rich regions of variable length.^1,19,27 The identification of h-fibroin gene sequences has been difficult because of their length (>20 kilobase pairs [kbp]) and their highly repetitive regions.^31,32,33 However, partial h-fibroin sequences of six species of Trichoptera were derived from sequencing the ends of cDNAs^{17,23,26,27,34} and the combination of long- and short-read sequencing approaches resulted in the assembly of two full-length h-fibroin sequences.^31,35 However, the h-fibroin could not be assembled in more than 20 species with these sequencing techniques. The lack of high-quality, full-length h-fibroin sequences has hindered the characterization and comparison of the primary molecular structure across trichopteran clades with different silk usage. Sequencing and assembling the entire repetitive central region of the h-fibroin is a crucial step toward understanding how phenotypes are encoded genetically,³² because the repetitive regions are responsible for the strength and elasticity properties of silk fibers. Recently, new genomic long-read sequencing techniques with low sequencing error rates (e.g., PacBio HiFi) and new genome assembly tools (e.g., hifiasm³⁶) allowed for full-length assembly of the h-fibroin gene sequences of four Trichoptera species (two case makers^32,33; one retreat and one cocoon maker³⁷).

In this study, we increased the number of high-quality full-length Trichoptera h-fibroin sequences from four to eleven. Specifically, we identified complete protein-coding h-fibroin gene sequences from seven high-quality genomes, including two newly assembled genomes and five publicly available genome assemblies. To increase the number of h-fibroin sequences of retreat and cocoon makers, we generated two de novo genomes and identified the h-fibroin sequences in these as well. Our final taxon sampling covers all three major silk usage strategies (six case makers, three retreat makers and two cocoon makers). We characterized and compared the primary structure of these h-fibroins and compared their amino acid composition to investigate differences between species with different silk usage. To examine differences between terrestrial and aquatic silk usage, we compared the amino acid composition of the h-fibroin to that of various terrestrial Lepidoptera.

Results

Expansion of genomic resources and heavy-chain fibroin sequences

Aquatic insects have been neglected concerning genome sequencing efforts³⁸ and the lack of well-resolved genome assemblies has hindered progress in understanding the genomic basis of aquatic insect traits, such as silk. Here, we used PacBio HiFi sequencing to generated genome assemblies for two species of Trichoptera: Leptonema lineaticorne (retreat-/capture net maker, 525,771 ccs-sequences with a total of 6,127,219,457 bp, ∼24.5× sequencing coverage) and Himalopsyche tibetana (cocoon maker, 1,391,005 ccs-sequences with a total of 14,730,644,101 bp, ∼23.3 × sequencing coverage). Genome size estimates derived from different methods were consistent. For L. lineaticorne, Genomescope2 revealed a genome size of 249,938,531 bp (http://qb.cshl.edu/genomescope/genomescope2.0/analysis.php?code=Tn1oE0FzenfoB8gl5Y0L). The back-mapping approach revealed a genome size of 265.85Mb (Figure S1). For H. tibetana, K-mer analysis estimated the genome size to 631,263,655 bp (http://qb.cshl.edu/genomescope/genomescope2.0/analysis.php?code=c2kegs2IOSe5C6OQiSpO). Back-mapping approach revealed a genome size of 683.7 Mb (Figure S3). Our de novo assemblies of L. lineaticorne and H. tibetana rank among the highest quality assemblies for Trichoptera for gene completeness (i.e., more than 96% complete BUSCOs, assembly lengths congruent with estimated genome size, Table 1) and contiguity (e.g., contig N50: ∼29 Mbp in H. tibetana; number of contigs: 65 in L. lineaticorne, Table 1). Re-mapping of the raw reads to the assembly revealed that 98.2% (H. tibetana) and 99.94% (L. lineaticorne) could be unambiguously placed with expected coverage distribution per position (Figures S1 and S3). BlobTools³⁹ detected no contaminations (Figures S2 and S4). We identified h-fibroin genes in the new L. lineaticorne and H. tibetana assemblies and in five previously published genome assemblies (Tables 2 and S1).

Table 1.

Comparison of currently published high-quality long-read-based genome assemblies of caddisflies

Species	Suborder (silk usage)	Accession number	Assembly length (bp)	Contig N50 (contig/scaffold)	No. of contigs/scaffolds)	% BUSCO (n = 2124)
Atopsyche davidsoni40	Basal Intergripalpia (cocoon)	GCA_022113835.1	370,818,532	14,095,054/n.a.	80/n.a	C:96.9%[S:96.1%,D:0.8%],F:1.8%,M:1.3%
Himalopsyche tibetanaa	Basal Intergripalpia (cocoon)	JAPJYX000000000	691,323,649	28,889,006/n.a.	282/n.a.	C:96.5%[S:95.6%,D:0.9%],F:2.4%,M:1.1%
Eubasilissa regina32	Integripalpia (tube case)	GCA_022840565.1	917,621,729	32,427,664/n.a.	123/n.a.	C:95.5%[S:94.8%,D:0.7%],F:3.0%,M:1.5%
Glyphotaelius pellucidus (DtoL)41	Integripalpia (tube case)	GCA_936435175.1	1,037,123,706	8,185,058/36,814,344	285/57	C:90.3%[S:89.5%,D:0.8%],F:6.8%,M:2.9%
Hesperophylax magnus33	Integripalpia (tube case)	GCA_026573805.1	1,215,205,050	11,205,906/n.a.	980/n.a.	C:91.4%[S:89.0%,D:2.4%],F:6.0%,M:2.6%
Limnephilus lunatus (DtoL)42	Integripalpia (tube case)	GCA_917563855.2	1,269,651,477	18,993,099/95,392,806	139/39	C:89.7%[S:88.9%,D:0.8%],F:7.4%,M:2.9%
Limnephilus marmoratus (DtoL)43	Integripalpia (tube case)	GCA_917880885.1	1,629,971,709	8,018,677/56,174,236	395/68	C:90.4%[S:89.3%,D:1.1%],F:6.7%,M:2.9%
Limnephilus rhombicus (DtoL)	Integripalpia (tube case)	GCA_929108145.2	1,578,808,083	10,796,652/54,234,467	272/62	C:89.8%[S:88.6%,D:1.2%],F:7.1%,M:3.1%
Leptonema lineaticornea	Annulipalpia (retreat)	GCA_024500535.1	273,010,349	13,827,090/n.a.	65/n.a.	C:96.1%[S:95.3%,D:0.8%],F:2.3%,M:1.6%
Cheumatopsyche charites44	Annulipalpia (retreat)	GCA_024721215.1	223,232,897	2,851,765/13,966,006	207/68	C:96.4%[S:95.9%,D:0.5%],F:1.9%,M:1.7%
Arctopsyche grandis37	Annulipalpia (retreat)	GCA_029955255.1	485,663,687	6,470,670/n.a.	676/n.a.	C:97.3%[S:93.5%,D:3.8%],F:1.6%,M:1.1%

C, complete; S, single; D, duplicated; F, fragmented; M, Missing.

^aThis study. BUSCO: % of complete BUSCOs is given based on BUSCO 5.2.2⁴⁵ using the endopterygota_odb10 dataset.

Table 2.

Full-length h-fibroins of 11 caddisfly species derived from long-read sequencing ordered by silk usage

Species	GenBank Accession	Silk usage	Gene length	CDS	Exon 1	Intron 1	Exon 2	Protein size	Molecular weight (kD)
Atopsyche davidsoni37	OQ787677	cocoon	23,701	23,637	67	61	23,570	7,878	755.16
Himalospyche tibetanaa	OQ983471	cocoon	20,892	19,233	36	1,656	19,197	6,411	628.28
Eubasilissa regina32	n.a., see32	case	25,256	25,161	40	92	25,121	8,386	815.86
Glyphotaelius pellucidusa	BK063450	case	24,412	23,592	39	817	23,553	7,864	801.63
Hesperophylax magnus33	OQ787679	case	26,666	25,872	42	791	25,830	8,624	875.02
Limnephilus lunatusa	BK063451	case	30,382	29,544	42	835	29,502	9,848	998.81
Limnephilus marmoratusa	BK063452	case	27,154	26,322	42	829	26,280	8,774	882.45
Limnephilus rhombicusa	BK063453	case	27,903	27,081	39	819	27,042	9,027	896.38
Leptonema lineaticornea	OQ983470	retreat	20,538	20,451	42	84	20,409	6,817	675.22
Cheumatopsyche charitesa	BK063449	retreat	18,745	18,669	42	73	18,627	6,223	633.908
Arctopsyche grandsis37	OQ787675	retreat	19,231	19,122	48	106	19,074	6,374	659.35

Gene length including introns and stop codon; CDS = protein coding DNA, gene length, CDS, exon, intron in bp; protein size in amino acids.

^aThis study.

Characterization of the primary structure of h-fibroin

Using the newly identified as well as the four previously published full-length, high-quality h-fibroin sequences,^32,33,37 our final taxon sampling for comparing the primary structure of h-fibroins covers the main clades of Trichoptera with different silk usage: fixed retreat-making Annulipalpia (n = 3), cocoon-making basal Integripalpia (n = 2), and tube case-making Integripalpia (n = 6). The structure of the h-fibroin gene across the 11 species revealed a similar organization of introns and exons. The h-fibroin gene was characterized by a short exon (36–48 bp) followed by a single intron (61–1,656 bp) and a long second exon (12,709–29,502 bp) leading to a total length of 18,745–30,382 bp.

The h-fibroin protein consists of non-repetitive n- and c-termini that were highly conserved across the different clades of caddisflies as well their sister order, Lepidoptera (B. mori). The n-terminus contained 105–117 residues (without the signal peptide, Figure 1A). There were 42.5% identical sites (% of columns in the alignment where all sequences are identical) and 74.3% pairwise identity (% of pairwise residues that are identical in the alignment, including gap versus non-gap residues, but excluding gap versus gap residues) among the caddisfly species. Alignment of the n-terminus of Lepidoptera B. mori to the n-termini of Trichoptera leads to a decrease in pairwise identity (65.4%) and identical sites (11.7%). The c-terminus consisted of 40 residues with 32.5% (10% when including B. mori) identical sites, and 65.3% (57.4% with B. mori) pairwise identity. A conserved cysteine was detected at position 19 in the c-terminus alignment (Figure 1B).

Figure 1.

Highly conserved regions of the h-fibroin

(A) n-terminus without signal peptide.

(B) c-terminus. Different silk usage is color-coded: violet: cocoon-, red: tube case-, blue: retreat-making, black: Bombyx mori (Lepidoptera, outgroup).

The terminal domains flanked a central region, composed entirely of repeating sequence blocks, which have been represented in several ways in the literature to describe the primary structure of h-fibroin.^27,30,46 For example, Frandsen et al.,³¹ represented each unique (SX)_nE motif as the beginning of a repeat. Structurally, this would correspond to defining each repeat as beginning with a single (SX)_nE ß-strand.¹⁷ In presenting the new h-fibroin sequences reported here, we have defined the repeating modules as each having two parts: first, a region comprising a variable number (1–7) of (SX)_nE motifs, each separated by short (8–24) stretches of intervening amino acids, and, second, a variable length (8–144 residues) G(glycine) - or G(glycine)-P(proline)-rich region. Schematically, in Figure 2, the two parts are represented by different symbols in each block. In the following paragraphs, we describe the repeating structural modules for each species analyzed in this study in detail. The full-length h-fibroin protein sequences are provided in Notes S1–S11. Schematic visualizations of each genus are presented in Figures S5–S13.

Figure 2.

Schematic visualization of the primary structure of the h-fibroin gene of one representative species per clade

Black boxes: n- and c-terminus, for sequences, see Figure 1. Gray boxes: transition regions (TR). The repeating modules (RM) have two parts which are represented by different symbols: A region comprising a variable number (1–7) of (SX)_nE motifs (red and orange circles correspond to red/orange residues in the sequence), each separated by short (8–24) stretches of intervening amino acids, and a variable length (8–144) G(glycine) - or G(glycine)-P(proline)-rich region (blue arrows correspond to blue residues in the sequence). Cocoon maker: Atopsyche davidsoni (AD): RM1_AD: Repeat module 1 occurs 81 times. Tube case maker: Glyphothaelius pellicudulus (GP): RM1_GP: Repeat module 1 and RM2_GP: Repeat module 2 occur 70 times each. Retreat maker: Leptonema lineaticorne (LL): RM1_LL: Repeat module 1 is repeated 14 times and RM2_LL: Repeat module 2 is repeated 13 times. The consensus sequence of each repeat module is given. The Glycine/Glycine-Proline-rich motifs vary in length. For full-length sequence see Notes S1–S11. Similar figures for each genus are given in Figures S5–S13. Phylogeny shown after.¹⁴ Figure created with BioRender.com.

Simple repeating structural modules in basal Integripalpia (cocoon makers)

The structure of the repetitive central region in cocoon-building A. davidsoni was simple compared to the other sequences investigated in this study (Figure 2, cocoon). The repeating module was embedded between two transition regions (TR), each of which occurs once and directly flanks the termini. In these transition regions, the sequence transitions from one type of module to another. The TRs had similarities to the repeating module but their sequence was unique, e.g., they occurred only one time in the h-fibroin and, in most cases, form a transition between the conserved termini and the repeating structural modules (Figure 2: cocoon, tube case) or between repeat modules (Figure 2: retreat). The h-fibroin of A. davidsoni included a single repeating module consisting of a (SX)₄E[15](SX)₃E region and a G-rich region of variable length (40–70 residues). The number in the square bracket refers to the short stretches of intervening amino acids. This module repeats 81 times across the sequence.

The repetitive region of H. tibetana was embedded between two transition regions (Figure S6). There were two repeating modules that are very similar. RM1 consists of a (SX)₆E[16](SX)₄E[15](SX)₆E[58](SX)₅E or (SX)₆E[16](SX)₄D[15](SX)₆E[58](SX)₅E (D: aspartic acid) motif and a glycine-rich motif of variable length (93–106 residues). RM2 was reduced to (SX)₆E[16](SX)₄E[15](SX)₆E. The glycine-rich motif was 56–90 residues long. RM1 occurred 26 and RM2 occurred 13 times.

Increasing structural complexity of repeating modules in Integripalpia (tube case makers) and Annulipalpia (fixed retreat makers)

The primary structure of the h-fibroin was variable within tube case-making basal Integripalpia. In all seven species sampled, the repetitive central region was flanked by two transition regions. However, the number and organization of repeat modules were diverse within this clade.

Specifically, in Eubasilissa regina which uses plant material (leaves) and silk to build a tube case, the h-fibroin had only one type of repeating module consisting of a (SX)₄E[8](SX)₄E[17](SX)₄E[11](SX)₃E[20](SX)₃E motif and a glycine-rich region (14–144 residues). It was repeated 39 times (Figure S7).

In Glyphotaelius pellucidula whose silk usage is similar to E. regina, the h-fibroin consisted of two repeat modules (Figure S8). RM1 consisted of a (SX)₄E[13](SX)₃E motif and glycine-rich region (38–81 residues), and RM2 consisted of a single (SX)₅E motif and glycine-rich region (12–51 residues). Each RM was repeated 70.

The h-fibroin of H. magnus (Figure S9), which uses stones the build tube cases, consisted of two repeat modules. RM1 contains a (SX)₅E[14](SX)₄E[11](SX)₄E motif and a glycine-rich region (41–81 residues) and occurred 69 times and was interrupted by 13 RM2, which consisted of a single (SX)₅E motif and a glycine-rich region (45 residues).

Within the three Limnephilus species, the primary structure of the h-fibroin was comparable (Notes S6–S8, Figure S10). This genus uses a diverse array of plant materials (wood, moss, and leaves) for tube case-making. There were two repeat modules. In L. lunatus, RM1 consisted of a (SX)₅E[14](SX)₄E[11](SX)₄E region and a glycine-rich region (14–82 residues). It occurred 96 times and was interrupted by ten RM2 which consisted of a single (SX)₅E motif and a glycine-rich region (14 residues), similar to G. pellucidula (Figure 2). In L. marmoratus, RM1 occurred 68 and RM2 17 times. In L. rhombicus RM1 occurred 59 and RM2 51 times.

In Annulipalpia (fixed retreat makers), in L. lineaticorne, the sequence was divided into two parts each with a distinct repeat module (Figure 2). RM1 consisted of a (SX)₄E[13](SX)₄E[11](SX)₄E[24](SX)₄E motif and glycine-rich region (68–92 residues) and is repeated 14 times. RM2 consisted of a (SX)₄E[14](SX)₄E[11](SX)₄E[19](SX)₃E[11](SX)₄E[24](SX)₄E[23](SX)₄E motif and a glycine-rich region (75–84 residues) and was repeated 13 times. There were two transition regions. TR1 in L. lineaticorne separated the two repeat modules and TR2 was located between RM2_LL and the c-terminus.

In Cheumatopsyche charites, the repetitive region was surrounded by two transition regions (Figure S13). Similar to L. lineaticorne, the h-fibroin was divided into two parts. The first part of the gene consisted of six repeat modules (RM1: SX₄E[20]-SX₄E and a glycine-proline-rich motif (112–141 residues): occurred 1x, RM2: (SX)₄E[15](SX)₄E[11](SX)₄E[25]-(SX)₄E[20](SX)₄E and a glycine-proline-rich motif (86–118 residues): occurred 2x, RM3: (SX)₄E[15](SX)₄E[11](SX)₄E and a glycine-proline-rich motif (44 residues): occurred 1x, RM4: (SX)₄E[15](SX)₄E[11](SX)₄E[25](SX)₄E[20](SX)₄E[20](SX)₄E and a glycine-proline-rich motif (86–118 residues): occurred 6x, RM5: (SX)₄E[15](SX)₄E[13](SX)₃E[16](SX)₄E[13](SX)₃E[18](SX)₃E[11](SX)₄E and a glycine-proline-rich motif (89 residues,: occurred 2 times, RM6: (SX)₄E[20](SX)₄E[20](SX)₄E and a G-P-rich motif (86 residues, occurred 1x). In the second part of the RM1 and RM3 are alternating 13 times.

In Arctopsyche grandis (Figure S12), the sequence included three internally repeating modules flanked by two transition regions. RM1 consisted of a (SX)₃E[12](SX)₄E[12](SX)₄E and a glycine-proline-rich motif (40–132 residues) and was repeated 24 times. RM2 contained a (SX)₃E and a glycine-proline-rich motif (22–63 residues). It occurred 31 times. RM3 comprised a (SX)₄E[12](SX)₃E and glycine-proline-rich motif (15–16 residues) and was represented 20 times.

Amino acid composition of the h-fibroin

Higher percentage of proline in the h-fibroin of fixed-retreat makers

In general, the amino acid composition of the protein sequence was conserved across the taxa that we sampled. Glycine and serine were consistently the most abundant residues across all three clades. However, despite these consistent patterns in composition, we observed some differences among clades. In retreat-making caddisflies, h-fibroin was characterized by a high amount of proline which ranged from 9.9 to 12.3% (n = 3). In contrast, the proportion of proline was much lower in the h-fibroin of tube case makers, ranging from 4 to 5.6% (n = 6). In the h-fibroin of cocoon-making caddisflies, the content of proline was even lower, ranging from 2.1–2.7% (n = 2).

Differences in the amino acid composition of h-fibroins of aquatic Trichoptera and terrestrial Lepidoptera

When comparing the amino acid composition of h-fibroins in Trichoptera to those of various terrestrial Lepidoptera (pyraloid moth,³² ermine moth,⁴⁷ bagworm,⁴⁸ silkworm,¹¹ butterfly⁴⁹), we find some consistent differences (Table 3). Although h-fibroins of both orders had high proportions of glycine (Trichoptera: 21.2–35.6%, Lepidoptera: 18.3–45.9%) and serine (Trichoptera: 9.3–17.2%, Lepidoptera:6.9–18.5%), h-fibroins in Lepidoptera had much more alanine (Trichoptera:0.1–4.9%, Lepidoptera: 21.9–40.52%). In addition, the Lepidoptera sequences exhibited a smaller percentage of charged residues. Negatively charged amino acids (aspartic acid and glutamic acid) ranged from 4–7.7% in the Trichoptera h-fibroin but only 1.1–2.4% in Lepidoptera. Positively charged amino acids (arginine, lysine) summed up to 7.6–16.9% in Trichoptera h-fibroins but were much lower in Lepidoptera (0.5–1.1%). Moreover, the amount of hydrophobic residues valine, leucine, and isoleucine was higher in the h-fibroins of caddisflies compared to Lepidoptera. Specifically, the amount of valine was 8.43% in retreat, 6.25% in cocoon, and 9.75% in case makers, but only 3.01% in Lepidoptera. Leucine ranged from 5.2% in retreat makers to 6.93% in case makers culminating in 10.1% in cocoon makers, whereas in Lepidoptera it was as low as 2.23%. The amount of isoleucine rages from 4.47% (case makers), 5.2% (retreat makers) to 7.3% in cocoon makers. H-fibroins of Lepidoptera only contained 1.36% of isoleucine.

Table 3.

Amino acid composition of full-length h-fibroins of Trichoptera and terrestrial sister order Lepidoptera

Residue	Trichoptera - retreat (n = 3)	Trichoptera - cocoon (n = 2)	Trichoptera - case (n = 6)	Terrestrial Lepidoptera (n = 4)
Gly (G)	23.60 (1.71)	33.55 (2.05)	29.43 (1.73)	31.14 (9.73)
Arg (R)	9.43 (1.59)	7.5 (1.5)	13.12 (1.79)	0.44 (0.11)
Lys (K)	1.67 (0.48)	1.3 (0.3)	2.12 (0.52)	0.32 (0.21)
His (H)	2.03 (0.97)	5.25 (1.75)	1.12 (2.23)	0.07 (0.04)
Ala (A)	4.20 (0.50)	1.75 (0.65)	0.35 (0.30)	29.00 (6.34)
Val (V)	8.43 (0.37)	6.25 (1.05)	9.75 (0.74)	3.01 (1.01)
Leu (L)	5.20 (0.71)	10.1 (0.7)	6.93 (1.27)	2.23 (2.28)
Ile (I)	5.20 (0.22)	7.3 (0.8)	4.47 (0.84)	1.36 (0.99)
Pro (P)	10.77 (1.09)	2.4 (0.3)	4.78 (0.69)	2.46 (2.02)
Ser (S)	11.77 (0.50)	12.1 (2.8)	15.20 (1.27)	13.66 (3.97)
Tyr (Y)	5.10 (1.36)	2.45 (1.35)	2.95 (1.05)	4.21 (1.61)
Asp (D)	3.37 (1.39)	1.5 (0.3)	2.25 (0.48)	0.61 (0.19)
Glu (E)	2.60 (0.14)	2.5 (0.1)	2.97 (0.18)	1.33 (0.57)
Thr (T)	2.90 (0.37)	3 (0.2)	1.62 (0.30)	0.91 (0.38)
Trp (W)	1.87 (0.57)	1.35 (0.25)	2.52 (0.82)	0.06 (0.08)

The mean is given for each amino acid and group of Trichoptera/Lepidoptera. The standard derivation is given in parentheses. Amino acid composition for each species is given in Additional data Files: Data S1–S17.

Discussion

In this study, we report eight new full-length h-fibroin sequences for caddisflies across a diverse set of silk usage. Two of these were generated from new genomic resources, whereas five were mined from previously published genomes,^41,42,43,44 highlighting the relevance of genome sequencing projects for the wider scientific community. The new full-length h-fibroins represent a substantial increase in the number of genomic resources available for the study of caddisfly silk and allowed us to compare the major silk gene for eleven species across the primary clades of Trichoptera, in which species exhibit different silk usage.

The genetic structure of h-fibroin of the eleven Trichoptera species showed similar organization of introns and exons, in line with previously reported h-fibroin sequences of Lepidoptera.^32,47 In addition to resolving the genetic structure, we compared the primary protein sequence of the h-fibroins. Consistent with previous research,^27,31 we found that structural elements of h-fibroin are conserved across Trichoptera. For example, of the species sampled, the n- and c-termini exhibit a high pairwise identity (n-terminus: 74.3%, c-terminus: 65.3%). In Lepidoptera, the conservation of the termini is linked with function. The n-terminus dimerizes and the c-terminus has been reported to interact with the light chain fibroin: the terminal cysteine forms an intermolecular disulfide bond with the light chain fibroin in the silkworm B. mori.⁴ Given similar patterns of conservation of the termini in Trichoptera, their function is likely conserved across both orders. The c-terminus of all Trichoptera was characterized by the presence of a cysteine at position 19 in the alignment (Figure 1). This suggests that disulfide crosslinking of fibroins occurs also in caddisflies and implies that covalent complex formation through the c-terminus is of similar importance for the structure, stability, and secretion of the fibroin complex as reported in B. mori.²⁸ In addition to the conservation of the termini, some conserved themes emerged in the central repetitive region, which consisted of repeating two-part structural modules, each containing a characteristic region of (SX)_nE motifs interspersed with glycine-rich (in the cocoon- and case makers) or glycine-proline-rich (in retreat makers) regions of variable length. The serines of the (SX)_nE motifs are extensively phosphorylated.^19,29,30 These phosphates then bind multivalent metal ions, which stabilize the silk and are responsible for the strength of the silk.^17,50,51 A structural model has been proposed, in which each (pSX)nE motif forms a β-strand, which in turn associates into anti-parallel Ca2+-stabilized ß-sheets.¹⁷ The β-sheets stack through alternating hydrophobic and Ca2+-phosphate interfaces creating microcrystalline β-domains. By this model, each repeating module would correspond to a structure in which anti-parallel β-sheets are linked with a glycine-rich or glycine-proline-rich spacer region. The (SX)_nE/glycine-rich blocks may combine to form a higher-order β-domain structure through intra- or intermolecular stacking of the [(SX)_nE]_m β-sheets. The crystalline β-domains would be separated by flexible and extensible glycine-/glycine-proline-rich regions. These signatures of conservation in repeat modules despite ∼280 million years of divergence¹⁴ and diverse silk usage, suggest a common mechanism for protein folding and silk formation across the underwater silks generated by all clades as suggested by.³¹

Despite the conservation in structure, we observed variation in the ordering and number of repeat modules. Our study builds on previous studies (i.e., ^{17,27,30,31,33,37,46}) by unveiling substantial diversity in the number and order of these repetitive structures. We observed a range of complexity in the repetitive structures across the phylogeny. In the simplest h-fibroin structure, sequenced from A. davidsoni, a single structural module was repeated 81 times (Figure 2). Slightly more complicated was the h-fibroin sequence of tube case-making G. pellucidula, which consisted of two repeat modules shuffled throughout the central region of the sequence (Figure 2). The sequences of the fixed retreat-makers were more variable, including L. lineaticorne, which was split into two parts, each of which consisted of a unique repeating module (Figure 2). The variations in repeating block patterns between the different clades may reflect adaptations for the diverse silk usage. For instance, cocoon-making H. tibetana and A. davidsoni are free-living as larvae and only produce pupal silk for building cocoons and pupal domes. The increase in h-fibroin sequence complexity scaled with silk usage diversity. The species in this study that construct both fixed retreats and capture nets exhibited the most variable h-fibroin sequences. The diversity in the number and order of the repeat modules may hold clues to unraveling the unique applications of silk across clades (e.g., case-, cocoon-making versus fixed retreat building) and may be directly responsible for differences in mechanical properties.

The amino acid composition of the h-fibroins was largely conserved across samples, with some notable exceptions. The proportion of proline was clade-specific. Although proline was found in low proportions in the h-fibroin of Integripalpia (cocoon (2.1–2.7%) and tube case makers (4–5.6%)), higher proportions were found in Annulipalpia sequences (9.9–12.3%). Annulipalpian larvae are generally characterized by their fixed retreats which serve as shelters. In addition, some annulipalpian families, including all of the species investigated in this study, also construct silken capture nets, which are used for capturing food and would presumably require more extensible silk. Future work in caddisflies should focus on linking the physical properties of the silk with variations in the h-fibroin sequence. For example, orb-weaver spiders use silk to build prey-capture spirals with high fiber extensibility. This is necessary to catch insects in flight without breaking the web. The abundance of proline content in the major ampullate spidroin MaSp2 of orb-weaver spider silk was linked to enhanced extensibility of these fibers^{52,53,54,55,56,57} because it increases the secondary structure disorder in the amorphous region.^52,53,54 Furthermore, Arakawa et al.⁵⁸ found that breaking strain was positively correlated with the presence of an amorphous, proline-rich region in the major ampullate spidroins MaSP1/2, which are often incorporated into dragline threads.⁵⁸

Despite similarities in the gene structure and conservation of the terminal regions of h-fibroin in aquatic Trichoptera and terrestrial Lepidoptera, we found consistent differences in the amino acid composition of the protein sequences. In general, to prevent limitations by protein content in their diets, non-essential amino acids glycine, alanine, and serine are the dominant residues in insect silk genes.² However, in contrast with Lepidoptera h-fibroins, which have high alanine content, Trichoptera h-fibroin sequences were extremely low in alanine. This is likely because of differences in how the proteins are folded. In Lepidoptera, the β-sheet structures are mainly derived from repetitive polyglycine-alanine domains (family Bombicidae, e.g., B. mori), (non-)/polyalanine-domains (family Saturniidae: saturniid moths), or a combination of both domains (bagworm family Psychidae⁴⁸) and, thus, these amino acids are important for the strength of the silk.^59,60 In contrast, as noted above, caddisfly h-fibroin β-sheets are primarily formed through the interaction of phosphorylated serine blocks with metal ions derived from their aquatic environment, an adaptation specific to aquatic dwelling species.⁵¹ H-fibroins of both, Trichoptera (9.3–17.2%) and Lepidoptera (6.89–18.5%) contain a high % of serines. Therefore, Ashton et al.²² suggested that one of the key molecular adaptations of a terrestrial ancestor silk to aquatic environments could have been kinase phosphorylating H-fibroin serines. We also detected a higher percentage of hydrophobic (Valine, Isoleucine, Leucine) and charged residues in caddisflies (Table 3), which has previously been hypothesized as an adaptation for aquatic silks.²⁷ Some previous research has been conducted on silk genes in aquatic insects.^61,62,63 For example, the aquatic black fly larva Simulium vittatum uses silk for larval adherence in rapidly flowing water and to construct pupal tents.⁶¹ Characterization of its silk gland proteins did not show sequence similarity to fibroins but revealed multiple phosphorylated serine residues and a high amount of hydrophobic and charged amino acids in the central region of the proteins.⁶¹ The authors argue that in an aquatic environment, hydrophobicity would lead to clumping and the greater proportion of charged amino acids might lead to proteins that are less hydrophobic compared to terrestrial silks.⁶¹ Previous studies on water-associated spiders also revealed higher concentrations of hydrophobic amino acid motifs (Glycine-Valine) in silk gene sequences (spidroins, 20–38%) compared to terrestrial spiders (2–4%). The spidroins of the semi-aquatic spiders showed little sequence similarity to fibroins of caddisflies and authors did not find evidence for similar serine-rich motifs.⁶² However, egg sacs of a semi-aquatic spider species contained calcium and phosphorus. These elements have not been detected in the egg sacs of non-aquatic spiders and might contribute to the water-repellant properties of the silk.⁶³

The new genomic data provided in this study was used to investigate the primary structure of h-fibroins across caddisflies. Although we observed conserved patterns in the primary structure of the h-fibroin, the amino acid composition and the number and arrangement of repeating modules varied among species with different silk usages. To understand the role of this variation in generating the myriad silk phenotypes that we observed across Trichoptera the sequences of the h-fibroin need to be linked with experimental evidence, such as mechanical testing as shown for spider silk.⁵⁸ Such studies are essential to gauge the potential of caddisfly silk in material science. The sequences that we have compiled here represent an important step toward performing such analyses.

Limitations of the study

This study provides characterizations of the primary structure of h-fibroin from a diverse set of caddisflies with different silk usage strategies (fixed retreat: n = 3, tube case: n = 6, cocoon builders: n = 2). We acknowledge that future studies including additional species of the three main clades of caddisflies (especially fixed retreat and cocoon builders) are important to underpin our findings. In addition, the study lacks comparisons with terrestrial Trichoptera (e.g., North American Philocasca demita or European Enoicyla pusilla) whose silk genes have not been sequenced to date. We, instead, compared the amino acid composition of the h-fibroin of Trichoptera with those of terrestrial species of sister order Lepidoptera (moths and butterflies). Despite these limitations, we hope that our study forms a foundation for future studies to screen and correlate amino acid motifs and other sequence features with quantifiable silk properties, such as mechanical measurements.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Biological samples
Leptonema lineaticorne tissue of whole body	This study	Sample ID EC20210919-2, NCBI Biosample ID: SAMN25408291
Himalopsyche tibetana tissue of thorax	This study	Sample ID BH20221015-01, NCBI Biosample ID: SAMN31697150
Critical commercial assays
Qiagen Genomic-tip extraction kit	Quiagen	NA
SMRTbell Express Prep kit v2.0	Pacific Biosciences, Menlo Park, CA	NA
Deposited data
Figshare deposited data for main text and supplemental analyses	This study	Figshare repository: https://figshare.com/s/03f88091eda258465d2b
GitHub Project: h-fibroin-visual including all custom-made scripts used in this study	This study	https://github.com/AshlynPowell/h-fibroin-visual
Leptonema lineaticorne sequence reads	This study	NCBI: Short Read Archive: SRR20711493
Leptonema lineaticornegenome assembly	This study	GenBank Accession: GCA_024500535.1
Leptonema lineaticorneh-fibroin	This study	GenBank Accession:OQ983470
Himalopsyche tibetanasequence reads	This study	NCBI: Short Read Archive: SRR22537910
Himalopsyche tibetanagenome assembly	This study	GenBank Accession: JAPJYX000000000
Himalopsyche tibetanah-fibroin	This study	GenBank Accession:OQ983471
Atopsyche davidsoni genome assembly	Ríos-Touma et al.36	GenBank Accession: GCA_022113835.1
Atopsyche davidsoni h-fibroin	Ríos-Touma et al.36	GenBank Accession: OQ787677
Eubasilissa reginagenome assembly	Kawahara et al.32	GenBank Accession: GCA_022840565.1
Glyphotaelius pellucidusgenome assembly	McSwan et al.48	GenBank Accession: GCA_936435175.1
Glyphotaelius pellucidus h-fibroin	This study	GenBank Accession:BK063450
Hesperophylax magnusgenome assembly	Hotaling et al.42	GenBank Accession: GCA_026573805.1
Hesperophylax magnus h-fibroin	Hotaling et al.42	GenBank Accession:OQ787679
Limnephilus lunatusgenome assembly	Austin et al.46	GenBank Accession: GCA_917563855.2
Limnephilus lunatush-fibroin	This study	GenBank Accession:BK063451
Limnephilus marmoratusgenome assembly	Clifford et al.47	GenBank Accession: GCA_917880885.1
Limnephilus marmoratus h-fibroin	This study	GenBank Accession:BK063452
Limnephilus rhombicusgenome assembly	WELLCOME SANGER INSTITUTE11	GenBank Accession: GCA_929108145.2
Limnephilus rhombicush-fibroin	This study	GenBank Accession:BK063453
Cheumatopsyche charitesgenome assembly	Ge et al.49	GenBank Accession: GCA_024500535.1
Cheumatopsyche charites h-fibroin	This study	GenBank Accession:BK063449
Arctopsyche grandisgenome assembly	Frandsen et al.39	GenBank Accession: GCA_029955255.1
Arctopsyche grandis h-fibroin	Frandsen et al.39	GenBank Accession: OQ787675
Software and algorithms
pbccs tool v6.6.0	NA	https://github.com/PacificBiosciences/pbbioconda
JELLYFISH v2.2.10	Marçais and Kingsford64	https://github.com/gmarcais/Jellyfish
GenomeScope 2.0	Ranallo-Benavidez65	http://qb.cshl.edu/genomescope/genomescope2.0/
Hifiasm v.0.13-r307	Cheng et al.66	https://github.com/chhylp123/hifiasm
BUSCO v5.2.2	Simão et al.67; Waterhouse et al.68	https://gitlab.com/ezlab/busco/-/releases/5.2.2
backmap.pl v0.5	Schell et al.69; Pfenninger et al.70	https://github.com/schellt/backmap
minimap2 v2.24	Li et al.71	https://github.com/lh3/minimap2
qualimap 2.2.1	Okonechnikov et al.72	http://qualimap.conesalab.org/
MultiQC v1.10	Ewels et al.23
bedtools v2.30.0	Quinlan et al.73	https://multiqc.info/
R v4.0.3	R Core Team 2021	https://www.r-project.org/contributors.html
BlobTools v1.1.1	Laetsch et al.43	https://github.com/DRL/blobtools
blastn 2.10.0+	Camacho et al.74	https://www.ncbi.nlm.nih.gov/books/NBK279690/
Geneious Prime 2022.1.1	NA	https://www.geneious.com/prime/
Augustus v.3.3.3	Hoff and Stanke73	https://bioinf.uni-greifswald.de/augustus/
SignalP 6.0	Teufel et al.74	https://services.healthtech.dtu.dk/service.php?SignalP
Muscle 3.8.425	Edgar et al.75	https://kbase.us/applist/apps/kb_muscle/MUSCLE_nuc/release
Expasy ProtParam	Gasteiger et al.76	https://web.expasy.org/protparam/
samtools v1.13	Li et al.71	https://github.com/samtools/

Resource availability

Lead contact

Further information can be requested via the lead contact, Jacqueline Heckenhauer (jacqueline.heckenhauer@senckenberg.de).

Materials availability

This study did not generate any new reagents.

Experimental model and subject details

Sample acquisition

We collected a single adult individual of L. lineaticorne in an Amazon Blackwater channel in Ecuador, Sucumbios, Sacha Lodge (Amazon Basin), Caño Anaconda (0°28′20.71″S; 76°27′59.08''W, elevation 237 m asl.) and a single larva of H. tibetana in Gasa, Bhutan (28°03.9477′N, 89°39.0019′E, elevation 3824 m asl.).

Method details

DNA extraction and whole-genome sequencing

We extracted high molecular weight from single individuals of L. lineaticorne and H. tibetana using the Qiagen Genomic-tip extraction kit and prepared DNA sequencing libraries following the instructions of the SMRTbell Express Prep. For each individual, one SMRT cell sequencing run was performed on the Sequel System II in CCS (circular consensus sequencing) mode using 30-h movie time.

Quantification and statistical analysis

Raw data processing, genome-size estimation, and whole-genome assemblies

We generated HiFi reads from the raw data (reads with quality above Q20) using PacBio SMRTlink software (https://github.com/PacificBiosciences/pbbioconda). We estimated genome size using sequencing reads and a k-mer-based statistical approach. After counting k-mers with JELLYFISH v2.2.10⁷⁷ using jellyfish count -C -s 25556999998 -F 3 and a k-mer length of 21 (-m 21) with the ccs-reads, we produced a histogram of k-mer frequencies with jellyfish histo. We ran GenomeScope 2.0⁶⁷ with the exported k-mer count histogram within the online web tool (http://qb.cshl.edu/genomescope/genomescope2.0/) using the following parameters: k-mer length = 21, Ploidy = 2, Max kmer coverage = 10000. We assembled the L. lineaticorne and H. tibetana genomes, with hifiasm v0.13-r307³⁶ with the default settings.

Assembly quality control

We evaluated the assembly quality based on continuity (QUAST v5.0.2⁷⁸) and completeness of Benchmarking Universal Single-Copy Orthologs (BUSCOs) with BUSCO v5.2.2^45,64 using the lineage dataset endopterygota_odb10 in genome mode. In addition, we calculated the back-mapping rate of the HiFi reads to the assemblies using backmap.pl v0.5^65,66 with the parameter -hifi. Other parameters were kept as default. This wrapper script automatically maps the reads to the assembly with minimap2 v2.24⁷⁹ and executes qualimap v 2.2.1,⁶⁸ MultiQC v1.10,⁶⁹ bedtools v 2.30.0,⁷⁰ and RScript v 4.0.3 (R Core Team 2021) to create the mapping quality report and a coverage histogram. In addition, it plots the coverage distribution and estimates of genome size from mapped nucleotides divided by the mode of the coverage distribution (>0). The final genome assemblies were screened for potential contaminations with taxon-annotated GC-coverage (TAGC) plots using BlobTools v1.1.1.³⁹ For this purpose, the bam file resulting from the back-mapping analysis was converted to a BlobTools readable.cov file with “blobtools map2cov”. Taxonomic assignment for BlobTools was done with blastn 2.10.0+⁷¹ using -task megablast and -e-value 1e-25. The blobDB was created and plotted from the cov file and blast hits. Contaminations of adapters detected by NCBI in the L. lineaticorne genome assembly were filtered out with samtools v1.13 faidx.⁷²

Identification and annotation of heavy-chain fibroins

Recently, the Wellcome Sanger Institute published four high-quality caddisfly genomes (Glyphotaelius pellucidus: GCA_936435175.1, Limnephilus lunatus: GCA_917563855.2, Limnephilus marmoratus: GCA_917880885.1, Limnephilus rhombicus: GCA_929108145.1) in the course of the Darwin Tree of Life Project. An additional high-quality genome of a retreat-making caddisfly was published by Ge et al.⁴⁴ We identified the h-fibroin genes in these assemblies by using tBLASTn to search the assemblies with the conserved n- and c-termini with query sequences from previously published species Hesperophylax sp.,¹⁷ Limnephilus decipiens AB214509²⁷ and Rhyacophila obliterata AB354689.1 and AB354588.1²³ in Geneious Prime 2022.1.1 (https://www.geneious.com) with default settings. After verifying that both BLAST hits (hit with n- and hit with c-terminus) were isolated to the same contig in the genome assembly, we extracted the sequences and 1,000 bp of flanking regions from the assembly using the sequence view “extract” in Geneious and annotated this region using Augustus v.3.3.3.⁷³ Introns that did not affect reading frames were manually removed from the annotation and h-fibroins were manually curated (see supplementary table). Protein coding nucleotide sequences were translated with the Geneious Tool “Translate” using the standard genetic code. We used the same approach to extract and annotate the h-fibroin in L. lineaticorne but using the termini of Parapsyche elsis.³¹ For H. tibetana we used termini of Rhyacophila obliterata. For details see Table S1. We predicted signal peptides and the location of their cleavage site of the h-fibroin protein sequences with the SignalP 6.0 server (https://services.healthtech.dtu.dk/service.php?SignalP,⁷⁴) using the following settings: organism = Eukarya, model mode = slow.

Comparison of heavy-chain fibroins across caddisflies clades

We used the previously published full-length h-fibroin sequences^32,33,37 and the newly identified h-fibroin sequences generated here to compare their primary structure. To compare conserved regions of the h-fibroin proteins, we aligned the n- and c-terminus each (without the signal peptide) in Geneious using the Muscle 3.8.425⁷⁵ plugin with a maximum of 1,000 iterations. We compared % pairwise identity and % of identical sites in Geneious. For each species, we used custom-made scripts to split the silk gene into repeat modules (https://github.com/AshlynPowell/silk-gene-visualization/tree/main). For schematic visualization of the primary structure, we generated a consensus sequence of all representative sequences of each repeat module by aligning these in Geneious using the Muscle 3.8.425 plugin with a maximum of 1,000 iterations. We used Expasy ProtParam⁷⁶(https://web.expasy.org/protparam/) to compute the molecular weight and the amino acid composition of each sequence. For comparison, we also calculated the amino acid composition of the four available Lepidoptera full-length h-fibroins [(silkworm Bombyx mori¹¹; Indianmeal moth Plodia interpunctella,³²; painted lady butterfly Vanessa carduii³⁷ and spindle ermine moth Yponomeuta cagnagella⁴⁷] using ProtParam. The amino acid composition of the bagworm Eumeta veriegata was obtained from Kono et al. 2019.⁴⁸

Published: July 4, 2023

Data and code availability

• Genomic data (sequence reads and assemblies) have been deposited to GenBank and are publicly available. For accession numbers, see key resources table. The data supporting the results of this article (all data associated with quality control of the assemblies, full-length h-fibroin gene sequences including introns, h-fibroin protein-coding nucleotide sequences and h-fibroin protein sequences newly identified in this paper) are available on Figshare https://doi.org/10.6084/m9.figshare.20407101. This paper utilizes existing, publicly available data. For accession numbers of these datasets see key resources table.

• All original code has been deposited at GitHub and is publicly available at: https://github.com/AshlynPowell/h-fibroin-visual

• Any other additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.