Natural Templates for Coiled-Coil Biomaterials ... - ACS Publications

Nov 8, 2012 - Convergently-evolved structural anomalies in the coiled coil domains of insect silk proteins. Tara D. Sutherland , Holly E. Trueman , An...
1 downloads 11 Views 2MB Size
Download PDF
Article pubs.acs.org/Biomac

Natural Templates for Coiled-Coil Biomaterials from Praying Mantis Egg Cases Andrew A. Walker,†,‡ Sarah Weisman,‡ Tsunenori Kameda,§ and Tara D. Sutherland*,‡ †

Research School of Biology, Australian National University, Canberra, Australia, 0200 Ecosystem Sciences, Commonwealth Scientific and Industrial Research Organisation, Black Mountain Laboratories, Acton, Canberra, Australia, 2601 § National Institute of Agrobiological Sciences, Tsukaba, Ibaraki, Japan, 305-8602 ‡

S Supporting Information *

ABSTRACT: Whereas there is growing interest in producing biomaterials containing coiled-coils, relatively few studies have made use of naturally occurring fibrous proteins. In this study, we have characterized fibrous proteins used by mother praying mantises to produce an extensive covering for their eggs called an ootheca and demonstrate the production of artificial ootheca using recombinantly produced proteins. Examination of natural oothecae by infrared spectroscopy and solid-state nuclear magnetic resonance revealed the material to consist of proteins organized predominately as coiled-coils. Two structural proteins, Mantis Fibroin 1 and Mantis Fibroin 2, were identified in ootheca from each of three species. Between species, the primary sequences of both proteins had diverged considerably, but other features were tightly conserved, including low molecular weight, high abundance of Ala, Glu, Lys, and Ser, and a triblock-like architecture with extensive central coiled-coil domain. Mantis fibroin hydrophobic cores had an unusual composition containing high levels of alanine and aromatic residues. Recombinantly produced mantis fibroins folded into coiled-coils in solution and could be fabricated into solid materials with high coiled-coil content. The structural features of mantis fibroins and their straightforward recombinant production make them promising templates for the production of coiled-coil biomimetics materials.

■

shown to modulate fiber thickness17 and branching frequency,18 and hydrogels based on coiled-coil fibrils have been developed.10,15,19 An alternative approach toward developing coiled-coil biomaterials is to employ naturally occurring coiled-coil fibrous proteins. Extracorporeal fibrous proteins such as silks have been extensively used as templates for biomaterials. In particular, biomaterials made from regenerated and recombinant portions of the β-sheet forming silk proteins from spiders and silkworms have been used for cell culture, drug delivery vehicles, tissue regeneration scaffolds, and biosensors.20,21 This approach has also been exploited successfully using recombinant silk proteins from honeybees22,23 and regenerated silk proteins from hornets24 that have high coiled-coil content.25−27 However, in comparison with β-sheet-forming silk proteins, the development of biomaterials based on coiled-coil silks is in its infancy, partially due to the paucity of suitable template proteins and knowledge of the natural processes by which they assemble. A praying mantis ootheca is an extensive structure of interconnected proteinaceous sheets that is produced by females to protect their eggs from predators and extremes of weather (Figure 1). In 1956, just 5 years after Linus Pauling

INTRODUCTION The coiled-coil protein motif has optical, mechanical, and biological properties that are distinct from other kinds of tertiary structures.1−3 In addition, the sequence-structure relationship of coiled-coils is one of the best understood of any protein motif:4 coiled-coils are formed by proteins with repeating heptad motifs of the form (abcdefg)n, where the first (a) and fourth (d) positions are generally occupied by more hydrophobic residues than other positions. In the helical conformation, the (a) and (d) residues form a hydrophobic stripe down one side of the helix, and individual helices associate to sequester these residues from the solvent, leading to the formation of a coiled-coil. Because the structure is reliant on amino acid character rather than identity, coiled-coilforming sequences show great diversity in their primary sequences. The flexibility of coiled-coil sequences and the existing knowledge of how they fold make them particularly amenable to rational engineering. For the above reasons, there is growing interest surrounding the inclusion of coiled-coil structures in biomaterials.5−8 Over the last two decades, much progress has been made in this regard, chiefly using de novo designed peptides. The selfassembly of coiled-coil fibrils in solution has been demonstrated using a variety of peptides that differ in oligomerization state, formation of blunt- or sticky-ended superhelices, and molecular architecture.9−16 Peptide sequence modification has been © 2012 American Chemical Society

Received: October 9, 2012 Revised: November 6, 2012 Published: November 8, 2012 4264

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

assembly during dehydration of a concentrated protein solution,33 has existing parallels in manufacture of materials from synthetic polymers.36 In this study, we have used infrared spectroscopy and solidstate nuclear magnetic resonance (NMR) to investigate the composition and molecular structure of ootheca. Primary sequences of ootheca proteins were identified by comparing cDNA sequences from ootheca-producing glands with peptide sequence information from tryptic digests of ootheca proteins. Ootheca protein primary sequences were then analyzed with regard to their secondary and tertiary structure using bioinformatic algorithms. To identify the functional features of ootheca proteins, we conducted analysis in parallel on three praying mantis species (Mantidae, subfamily Mantinae): the false garden mantis Pseudomantis albof imbriata, the purplewinged mantis Tenodera australasiae, and the brown stick mantis Archimantis monstrosa. We demonstrate that it is possible to express ootheca structural proteins in E. coli and to use recombinant proteins to fabricate films and scaffolds with high coiled-coil content.

■

MATERIALS AND METHODS

Insects and Dissections. Mantises were collected from the Canberra region (ACT, Australia), kept at room temperature in large glass jars, and fed field crickets (Acheta domestica, Pisces Live Food). Female collaterial glands were dissected in phosphate-buffered saline, pH 7.0, and stored in RNAlater (Ambion), at −80 °C. Fourier Transform Infrared Spectroscopy. Spectra were obtained on a Bruker Tensor 37 Fourier transform infrared (FTIR) spectroscope with a MIRacle diamond ATR attachment using OPUS software. Averages of 128 scans were recorded at a resolution of 2 cm−1. Solid-State Nuclear Magnetic Resonance. High-resolution solid-state 13C NMR spectra were recorded using an Avance 600 WB (Bruker), with a magnetic field of 14.1 T. The spectrometers were operated at a 13C NMR frequency of 150.94 MHz. Samples were placed in a solid-state probe and spun at a magic-angle spinning (MAS) frequency of 10.0 kHz in a 4.0 mm φ zirconia rotor. For the cross-polarization (CP) experiments, the 1H 90° pulse length of 3.5 μs, 1 H−13C CP contact of 70 kHz, and high-power 1H decoupling by the spinal64 method were employed. The repetition time for CP experiments was set at 3.0 s. All spectra were calibrated using adamantine as the standard, and the chemical shift of the adamantine CH 2 peak appearing at 29.5 ppm was referenced to the tetramethylsilane (TMS) peak appearing at 0 ppm. Construction of cDNA Libraries. Total RNA was prepared from the primary collaterial glands of a single gravid female praying mantis from each of the three species investigated using RNAqueous-4PCR (Ambion). mRNA was isolated from total RNA using Micro-FastTrack 2.0 (Invitrogen). Libraries of cDNAs were constructed using a Cloneminer II cDNA Library Construction kit (Invitrogen) according to the manufacturer’s instructions. DNA was isolated from individual cDNA clones using a Qiaprep Spin Miniprep kit (Qiagen) and the size of inserts was determined by restriction digest and gel electrophoresis. Clones containing cDNA inserts larger than 300 bp were sequenced from the 5′ end by Micromon Services (Monash University, Melbourne, Australia) using an Applied Biosystems 3730S Genetic Analyzer and Applied Biosystems PRISM BigDye Terminator Mix cycling chemistry. Protein sequences encoded by the cDNA were identified by the presence of an open reading frame longer than 300 bp or a poly-A tail longer than 15 continuous nucleotides at the 3′ end of the clone. A library of protein sequences encoded by cDNA sequences was generated in silico using EMBOSS Transeq.37 All cDNA sequences were submitted to Genbank. Mass Spectroscopy. Solubilized ootheca was prepared by collecting wet ootheca material during oviposition and immediately placing it in 6 M guanidinium hydrochloride containing protease

Figure 1. Praying mantis ootheca. (A) Female mantis in the process of fabricating ootheca. Credit: Photo provided by Indrani Ghose. (B) A mature ootheca after hardening. (C) Scanning electron micrograph of protein sheets inside ootheca. Scale bar is 400 μm.

proposed the existence of the α helix, K. M. Rudall conducted X-ray scattering experiments on ootheca and proposed them to consist of proteins in coiled-coil conformation.28 Models have been subsequently been proposed for the structure of ootheca crystallites, in which dimeric coiled-coils are arranged in layers with adjacent coiled-coils staggered with respect to each other.28−30 Ootheca proteins are reported to have molecular weights between 43 and 60 kDa29−31 and to have a balanced amino acid compositionincluding Glx (21%), Ala (15%), Lys (10%), and Asx (10%)with no nonstandard amino acids.32 The proteins are produced and stored at high concentrations in a dedicated collaterial gland,33 then deposited as an aerated aqueous solution which hardens into a tough porous matrix.28,32 Mantis ootheca, like hymenopteran silk, is a solid coiled-coil material fabricated from a concentrated protein solution. Nonetheless, the independent evolutionary origins of these two materials are reflected in fundamental differences in coiled-coil structure and fabrication pathways: whereas hymenopteran silk proteins form tetramers,34 mantis ootheca has been proposed to consist of dimeric coiled-coils.28−30 Whereas hymenopteran silk fibers are produced by drawing ordered protein solutions through an aperture called a spinneret,11 the protein solution used to make ootheca is deposited into the environment, and proteins self-assemble into ordered crystallites as the solution dries.28,32 The features listed above suggest ootheca proteins as suitable template proteins for biomaterials production. Because of their coiled-coil structure, ootheca proteins are likely to be amenable to rational sequence engineering without disrupting their overall structure. Their small size and diverse composition of standard amino acids suggest that the proteins could be produced in recombinant expression systems without encountering the common problems associated with producing fibrous proteins such as collagen and spider silks.35 Lastly, the fabrication process of natural oothecae, which involves self4265

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

Circular Dichroism. Far ultraviolet CD spectra of fibroin solutions (4% w/v) between 180 and 300 nm were measured using a 0.2 mm path-length cell (20/O/Q/0.2, Starna) in a Chirascan spectrometer (Applied Photophysics) at 20 °C. Spectra were collected between 180−300 nm with 1 nm increments and an integration time of 0.5 s. For each final spectrum, three spectra were averaged and a blank subtracted. Fabrication of Solid Protein Materials. To prepare films, we placed 2 mL of protein solutions (4% w/v) in flat plastic dishes and allowed them to air-dry at room temperature. To prepare sponges, the same protein solutions were poured in silicone rubber molds (14 × 5 × 6 mm; RL060, ProSciTech), frozen at −20 °C overnight, and placed in a freeze-dryer (FD355DMP, FTS Systems) for 24 h to generate sponges of 13 × 5 × 5 mm, which were stored at room temperature.

inhibitors (Complete mini EDTA-free, Roche), 5 mM EDTA, and 2mercaptoethanol. Denaturant solution was removed from solubilized samples by repeated concentration with a Centricon-10 centrifugal filter device (Millipore) and dilution with 0.1 M sodium dodecyl sulfate (SDS). Solubilized ootheca proteins were separated by SDSpolyacrylamide gel electrophoresis (PAGE) using NuPage 4−12% BisTris gels and 2-(N-morpholino)ethanesulphonic acid running buffer (Invitrogen) and stained with Coomassie Brilliant Blue (SigmaAldrich). Liquid chromatography/mass spectroscopy (LC/MS) was performed on protein separated by SDS-PAGE and then digested with trypsin (Sigma-Aldrich) as previously described.26 Mass spectral data sets were analyzed using Spectrum Mill software (Agilent). Sequence Analysis. Sequences were aligned and manipulated using BioEdit.38 Signal peptides were predicted using the SignalP 4.0 algorithm.39 Protein sequences were aligned using CLUSTALW.40 Secondary structure was predicted using PSIPRED.41 Coiled-coil domains were predicted using MARCOIL.42 Residue hydrophobicity was calculated using octanol-to-water solvation energies.43 Confidence levels cited for trends in amino acid character were calculated using Student’s one-tailed unpaired t test. Recombinant Protein Expression and Purification. Recombinant expression constructs (without the signal peptides) were made by PCR amplification of full-length sequences encoding mantis fibroins. The primers used were as follows, with Nco1 and BamH1 sites underlined, and coding sequences italicized: T. australasiae Mantis Fibroin 1, GGA ATT CCC ATG GGC TCT CCC TTG GAA GAC AAA TAC and C GGC GGA TCC TTA TTA CAG ACC TTC GCC GGA AC; T. australasiae Mantis Fibroin 2, GGA ATT CCC ATG GGC AAG AAA CAT GAAG TAA TGA and C GGC GGA TCC TTA TTA TCC GTG GTA GTT GGA GTG G; A. monstrosa Mantis Fibroin 1, GGA ATT CCC ATG GGC TCT CCC TTG GAA GAA AAA TAT G and C GGC GGA TCC TTA TTA ACT CAT TCC TTC ACC TTC AGT T; A. monstrosa Mantis Fibroin 2, GGA ATT CCC ATG GGC AAG AAA CAC GAA GCA and C GGC GGA TCC TTA TTA TGC TCC GTG GTA GTT GGA; P. albof imbriata Mantis Fibroin 1, GGA ATT CCC ATG GGC TCA CCC TTG GAA GAA AAA TAT and C GGC GGA TCC TTA TTA TTC ATC GCC GTA AGA CAT TT; P. albofimbriata Mantis Fibroin 2, GGA ATT CCC ATG GGC AAG AAT CAC GAA GTA ATG and C GGC GGA TCC TTA TTA TGC TCC GTG GTA GTT GGA G. PCR amplicons were then digested with Nco1 and BamH1 (New England Biosciences) and cloned into pET14b vector (Novagen) using T4 DNA ligase (New England Biosciences). Successful cloning was verified by restriction digest, and DNA sequencing. E. coli Rosetta2 DE3 cells (Novagen) were transformed with the ligated vectors and allowed to grow on LB plates containing 100 μg/mL ampicillin and 34 μg/mL chloramphenicol overnight. Colonies from plates were used to inoculate a 50 mL starter culture in a 250 mL flask grown for 20 h, which was used to inoculate a 500 mL culture in a 2.5 l flask grown for 24 h. Both cultures were grown in Overnight Express Autoinduction media (Merck) on a 200 rpm shaker at 37 °C. Cultures were pelleted by centrifugation at 13 000g for 10 min and stored at −80 °C for later use. To isolate inclusion bodies, we disrupted pellets by adding 5 mL of BugBuster Master Mix (Novagen) per 1 mL of wet cell paste, applying a homogenizer for 15 s, and agitating on an axial rotator for 15 min at 4 °C. Inclusion bodies were pelleted by centrifugation at 13 000g for 10 min and subjected to three further washes with the Bugbuster reagent. Pellets were resuspended by adding 2% SDS and agitating gently for 16 h at 4 °C. BCA protein assay kits (Pierce) were used to determine the concentration of the protein in 2% SDS and used to calculate expression levels. Detergent was then removed by dialyzing against 100 mM KCl at 4 °C for 16 h using a 3.5 kDa MWCO Slide-ALyzer dialysis cassette (Pierce) and centrifuging at 13 000g for 5 min to remove the insoluble potassium salt of dodecyl sulfate. Final protein solutions for fabrication of solid materials and circular dichroism (CD) experiments were prepared by dialyzing against 50 mM NaCl for 16 h at 4 °C using a 3.5 kDa MWCO Slide-A-Lyzer dialysis cassette (Pierce) and were stored at 4 °C. Mixtures of mantis fibroin 1 and mantis fibroin 2 were prepared by mixing equal quantities of the two proteins before removal of SDS.

■

RESULTS

Oothecae Are Composed of Low-Molecular-Weight Proteins Folded into Coiled-Coils. Attempts to solubilize oothecae collected from the wild using concentrated solutions of lithium bromide, guanidinium hydrochloride, or SDS were unsuccessful, even in the presence of reducing agents and at 85 °C. However, freshly produced material collected from animals in culture during oviposition could be dissolved in guanidinium hydrochloride on the condition that it was collected before the material hardened. Solubilized ootheca proteins were analyzed by SDS-PAGE under reducing conditions to reveal the number and molecular weight of proteins present in the material. This procedure was carried out for P. albof imbriata and T. australasiae, resolving five to eight protein bands, with the most abundant proteins migrating between the 40 and 60 kDa protein standards (Figure 2). During the period of this study, we did not observe A. monstrosa females engaged in ootheca production, so no soluble ootheca could be collected. The FTIR spectroscopy and solid-state NMR spectra of oothecae from each of the three study species were similar, indicating that they had essentially the same molecular structure. No non-protein organic components were detected using either technique. The amide I and II regions of FTIR spectra showed broad peaks with maxima at 1643−1647 and 1537−1539 cm−1 (Figure S1 of the Supporting Information). These peak positions are characteristic of coiled-coils but distinct from isolated α-helices, which have amide I maxima above 1650 cm−1.27,44 NMR spectra showed strong peaks at 176.3 and 54.1 ppm, indicating a high proportion of α-helical content and a low proportion of β-sheet (Figure 3, Table 1; 24). The peak attributed to Ala Cβ occurred between 16.2 and 16.7 ppm, depending on the species. The position of this peak is similar to that of other coiled-coil proteins such as keratin45,46 and αtropomyosin.47 Peaks at 116, 129, and 137 ppm were attributed to aromatic residues.45 Two Main Structural Proteins Identified by LC-MS. For each of our three study species, a cDNA library was constructed from the primary collaterial gland of a single gravid female. More than 60 cDNA inserts of greater than 300 bp were sequenced from each of the P. albofimbriata, T. australasiae, and A. monstrosa libraries (Table S1 of the Supporting Information). On inspection of the libraries, two cDNAs were notable because of their high abundance and because orthologous sequences occurred in each of the three species. Most cDNAs were detected a single time (Table 2), but the two common cDNAs combined made up 40−80% of each cDNA library (Table 2). To investigate whether the proteins encoded by the 4266

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

Figure 2. Oothecae proteins separated on a SDS PAGE gel. Left lane: P. albof imbriata; right lane: T. australasiae. Molecular weight markers corresponding to the Benchmark Protein Ladder (Invitrogen) are shown on the right.

Figure 3. Solid-state 13C NMR spectra from oothecae of three species of praying mantis: (A) P. albof imbriata, (B) T. australasiae, and (C) A. monstrosa.

Table 1. Structural Assignments for NMR Peaks from the Spectra of Praying Mantis Oothecae

highly abundant cDNAs were present in mantis ootheca, we used liquid chromatography−tandem mass spectrometry to identify individual protein bands from solubilized ootheca (Figure 2) by comparing them to in silico translated cDNA libraries. LC-MS experiments were carried out for P. albofimbriata and T. australasiae. The most intensely stained protein bands were identified as the products of the two highly abundant cDNA sequences, which we named Mantis Fibroin 1 and Mantis Fibroin 2 (accession numbers and LC-MS data in table 2). The intense band at 48 kDa in the protein gel of P. albofimbriata ootheca was found to contain both Mantis Fibroin 1 and Mantis Fibroin 2, whereas the intense bands at 50 and 44 kDa in the protein gel of T. australasiae were found to correspond to Mantis Fibroin 1 and Mantis Fibroin 2, respectively (Table 2). Three further P. albofimbriata proteins and five further T. australasiae proteins were identified by LC-MS. Because these proteins correspond to less intensely stained bands on protein gels of solubilized ootheca and are represented in the cDNA library by fewer sequences (Table 2), they are likely to be present in ootheca at lower levels than the two mantis fibroin proteins. Three of these minor proteins had sequence homology to proteins in Genbank’s nonredundant protein database (Table 2): two to insect cuticular proteins (multiple

peak position (ppm) P. albof imbriata

T. australasiae

A. monstrosa

16.3 21.2

16.7 21.4

16.2 20.6

25.2 36.2 40.5 54.1 56.3 116.4 129.4 136.9 157.1 176.2

24.8 36 40.5 54.1 56.5 116.4 129.2 136.5 157.1 176.3

25.2 35.8 40.8 54.1 56.4 116.1 129.2 136.2 156.9 176.3

assignment Ala Cβ (α-helix) Ala Cβ (β-sheet), Val and Thr Cβ side-chain carbons side-chain carbons Gly Cα Ala Cα (α-helix) various Cα aromatic residues aromatic residues aromatic residues Arg guanidine C carbonyl (α-helix)

hits with E < 10−10) and one to catalase proteins (multiple hits with E < 10−5). Mantis Fibroins Form Coiled-Coils with an Unusual Alanine/Aromatic Core. The features of Mantis Fibroin 1 and Mantis Fibroin 2 are summarized in Table 3 and shown in 4267

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

Table 2. Identification of Praying Mantis Egg Case Proteins sequence Pseudomantis albof imbriata Fibroin 1 (JQ421307) Fibroin 2 (JQ421308, JQ421309) JQ421298 JQ421299 JQ421300 Tenodera australasiae Fibroin 1 (JQ421310) Fibroin 2 (JQ421311, JQ421312) JQ421303 JQ421301 JQ421305 JQ421302 JQ421304 JQ421306 Archimantis monstrosa Fibroin 1 (JQ421313, JQ421314) Fibroin 2 (JQ421315) a

abundance in library (%)

SpectrumMill score (no. of peptides)a

24.6 24.6

78.1 (5) 119.2 (10)

10.5 3.5 1.8

71.1 (6) 79.0 (7) 25.2 (2)

31.5 13.1 5.4 4.3 4.3 3.3 2.2 1.1

database homology

protein band intensityb

MW on gel (kDa)b

predicted MWc

nd nd

high high

48 48

41 40

nd yesd nd

very low low ?e

30 58 48

nc nc nc

226.9 (16) 96.3 (8)

nd nd

high high

50 44

41 40

45.8(4) 127.4 (10) 149.1 (12) 94.6 (8) 23.4 (2) 47.3 (3)

nd yesd nd nd nd yesf

very low very low very low very low ?e medium

25 25 30 33 50 58

21 22 nc 29 41 nc

40.7

nd

38.4

41

nd b

41

c

d

e

SpectrumMill scores >20 are considered a confident identification. See Figure 1. Without signal peptide. Cuticular protein. Coincides with one or both of the mantis fibroin proteins on SDS-PAGE gel. fCatalase, nd: not detected; nc: could not be calculated because only a partial sequence was available.

Table 3. Comparison of the Features of Mantis Fibroin 1 and Mantis Fibroin 2 signal sequence length (residues)* molecular weight (kDa)* amino acid composition* predicted α-helix coiled-coil domain

Mantis Fibroin 1

Mantis Fibroin 2

22 residues 361−365 40.7−41.3 Ala (12.9−15.9%) Glu (11.0−12.7) Ser (8.5−10.5%) 72−81% 21−23 heptads

18 residues 342−351 39.7−40.5 Ala (10.2−13.9%) Glu (13.1−13.5%) Lys (9.7−10.2%) 76−78% 18−31 heptads

Supporting Information). The residue composition of these positions was highly unusual: approximately half of the residues in the (a) position of both proteins were alanine (Figure 5), which results in a much shorter average side-chain length at this position than the side-chain length in other positions (p < 0.03; Figure S2 of the Supporting Information). The (d) position of the core contained a low proportion of alanine residues, instead owing its hydrophobicity to a high proportion of aromatic residues (Figure 5), principally tyrosine. Recombinant Fibroins Form Coiled-Coils in Solution and Solids. Mantis Fibroin 1 and Mantis Fibroin 2 from each of the three study species were individually expressed in a standard recombinant expression system, E. coli Rosetta2 cells. Initially, all constructs were expressed at 28 and 37 °C, and inclusion bodies and the soluble fraction were examined by SDS-PAGE. For all constructs, the best expression was achieved over 24 h at 37 °C, with proteins being recovered from the inclusion bodies after solubilization in 2% SDS (Figure 6A). Detergent was then removed using a previously published method23 that uses KCl to precipitate dodecyl sulfate ions, which are then removed by centrifugation. Under these conditions, 40−400 mg/L of protein at high purity could be recovered per liter of expression (Figure 6A). Solutions of recombinant P. albofimbriata fibroins were individually examined by CD spectroscopy. The CD spectra from both solutions showed two minima at 209 and 220 nm

Figure 4. Both proteins begin with a signal peptide, as expected for proteins that are secreted into a gland lumen. The algorithm PSIPRED predicts that 72−81% of each mantis fibroin sequence folds into α-helices, and MARCOIL predicts that around 126−213 residues from each protein contribute to coiled-coils (37−61% of total sequence). Despite having many similar features (Table 3), the primary sequences of Mantis Fibroin 1 and Mantis Fibroin 2 could not be convincingly aligned. Within homologues, the signal peptide was strongly conserved (82−89% identity), the coiled-coil region showed moderate conservation (32−41% identity), and the terminal sequences were less conserved, except for a 24 residue region at the C-terminus of Mantis Fibroin 2 that was strongly conserved (Figure 4B). For each of Mantis Fibroin 1 from A. monstrosa, Mantis Fibroin 2 from P. albof imbriata, and Mantis Fibroin 2 from T. australasiae, we recorded two groups of sequences differing by one to two single nucleotide polymorphisms in approximately equal proportions (Table S2 of the Supporting Information). Because each cDNA library was constructed from a single female, these likely correspond to transcripts from maternal and paternal alleles. Within the coiled-coil domains of both Mantis Fibroin 1 and Mantis Fibroin 2, residues predicted by MARCOIL as positions (a) and (d) were significantly more hydrophobic than residues in other positions, as expected (p < 0.05; Figure S2 of the 4268

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

Figure 4. Primary sequence alignment of mantis fibroins with primary, secondary, and tertiary structural annotation. Panel A: Mantis Fibroin 1; Panel B, Mantis Fibroin 2. Pal = P. albof imbriata; Tau = T. australasiae; Amo = A. monstrosa. Predicted signal sequences are underlined. Amino acid identity (asterisk) or similarity (colon) between orthologues is indicated. Structural predictions for the P. albof imbriata sequences are shown above the sequences, with gray boxes indicating α-helical structure, black boxes indicating coiled-coil domains, and italics indicating predicted heptad positions.

(Figure 6B), a characteristic feature of α-helical proteins.1 The ratio of the ellipticity at 220 nm to the ellipticity at 209 nm can discriminate noninteracting α-helices from coiled-coils, with ratios exceeding unity indicating coiled-coils and ratios less than 0.91 indicating isolated helices.1 The measured ratios for Mantis Fibroin 1 and Mantis Fibroin 2 solutions were 1.01 and 1.03, respectively, indicating that each recombinant protein in isolation was able to form coiled-coils in solution. Films and sponges were fabricated from the coiled-coil solutions of recombinant P. albofimbriata Mantis Fibroin 1 or Mantis Fibroin 2 (Figure 6C). The FTIR spectra of these materials were very similar to spectra from native P.

albof imbriata ootheca (Figure 6D) with broad amide I peaks close to 1645 cm−1, the expected position for coiled-coils.

■

DISCUSSION In search of proteins suitable for the development of coiled-coil biomaterials, we studied the extensive oothecae produced by praying mantises (Figure 1). We used solid-state NMR and infrared spectroscopy to demonstrate that ootheca is made from proteins in a coiled-coil conformation, in agreement with previous X-ray and electron scattering data.28−30 The proteins from which these structures are formed were identified in each of three species using a combined proteomic and transcriptomics approach. cDNA libraries constructed from the 4269

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

Fibroin 1 and Mantis Fibroin 2 constitute the main structural proteins present in oothecae. Mantis Fibroin 1 and Mantis Fibroin 2 are very similar in length (Table 3), architecture (Figure 4), overall amino acid composition (Table 3), and core composition (Figure 5), suggesting these to be features suited to the fabrication of a solid coiled-coil material from a concentrated protein solution. In agreement with this notion, some strikingly similar features have convergently evolved in coiled-coil-forming silk proteins of hymenopterans. Mantis and hymenopteran fibroins both have tightly conserved lengths of 30−50 kDa and a triblock-like architecture with a central coiled-coil domain that occupies 37− 61% of the sequence flanked by termini that form other structures (Figure 4; 27). Compared with mantis fibroins, hymenopteran silk proteins have higher overall hydrophobicity and lack aromatic residues in their core. The differences between mantis and hymenopteran fibroins may relate to differences in preferred oligomerization state or differences in the fabrication processes they undergo. Because it was not possible to convincingly align Mantis Fibroin 1 and Mantis Fibroin 2 at the primary sequence level, it is not clear whether the two proteins are highly diverged paralogues or unrelated proteins that have convergently evolved similar features. Whereas the similarities between the two proteins suggest that they are functionally redundant, we observed that in each of the species examined genes encoding Mantis Fibroin 1 and Mantis Fibroin 2 were conserved and expressed at approximately equal levels. If the proteins were

Figure 5. Nature of amino acids at positions within the mantis fibroin coiled-coil. Data shown are pooled from three species. (A) Mantis Fibroin 1 and (B) Mantis Fibroin 2.

glands dedicated to ootheca production were dominated by transcripts from two genes, Mantis Fibroin 1 and Mantis Fibroin 2 (Table 2). The proteins encoded by these two genes were the most abundant proteins in oothecae (Figure 2). Both proteins were predicted to form coiled-coils, matching the experimental data collected from native oothecae. We conclude that Mantis

Figure 6. Recombinant expression, refolding, and material fabrication of mantis fibroins. (A) Purified mantis fibroins after expression in E. coli. Lanes 1 and 2: T. australasiae Mantis Fibroin 1 and Mantis Fibroin 2; Lanes 3 and 4: A. monstrosa Mantis Fibroin 1 and Mantis Fibroin 2; and Lanes 5 and 6: P. albofimbriata Mantis Fibroin 1 and Mantis Fibroin 2. Molecular weight markers corresponding to the BenchMark Protein Ladder (Invitrogen) are shown on the left. (B) Circular dichroism of recombinant P. albofimbriata mantis fibroin solutions. The dashed line shows data for Mantis Fibroin 1 and the solid line shows data for Mantis Fibroin 2. (C) Sponge made by freeze-drying P. albofimbriata Mantis Fibroin 2. Inset shows entire sponge, 12 mm long. Scale bar is 60 μm. (D) FTIR spectra of materials made from recombinant mantis fibroins. Solid black line: film made from P. albofimbriata Mantis Fibroin 2. Dashed line: sponge made by freeze-drying the same solution. The solid gray line shows the spectrum from a native P. albof imbriata ootheca. 4270

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

can be significantly modified while retaining their ability to fold and function correctly. This sequence flexibility, combined with high-level E. coli expression, opens up a broad range of possibilities for engineering the proteins to generate materials with non-native functionality while retaining the ability to produce a durable structural material. The combination of straightforward biosynthesis and potential for further engineering make mantis fibroins promising candidates for the development of coiled-coil biomaterials.

functionally redundant, then it would be expected that gene deletion or duplication events would be observed between species. Instead, the conservation and expression of both genes suggests that each has a distinct and critical role in ootheca formation. In any case, the purified recombinant forms of Mantis Fibroin 1 and Mantis Fibroin 2 are each capable of folding into coiled-coils and forming solid materials in the absence of the other (Figure 6). The predicted coiled-coil domains in the mantis fibroin sequences have an unusual composition of amino acids in their hydrophobic core. In contrast with the cores of classic coiledcoils that are dominated by residues with large aliphatic side chains,4 mantis fibroin cores feature high proportions of Ala in the (a) position and aromatic residues in the (d) position. Despite having the highest helix-forming propensity of all amino acids, Ala is generally regarded as an unfavorable core residue because of its low hydrophobicity and small size relative to residues such as Leu and Ile. However, Ala is common in the core of longer coiled-coil domains, including those of αtropomyosin48 and the hymenopteran silk proteins,27 where its reduced hydrophobicity is compensated by the extensive length of the coiled-coil domain. Aromatic amino acids are also regarded as unfavorable core residues because they are too bulky to pack favorably.4 Whereas there are several examples of artificial coiled-coils featuring aromatic residues in the core,49,50 regular inclusion of aromatics is uncommon in natural proteins. Although the pattern of residues we observed in the core of mantis fibroins is atypical, we demonstrated using CD and infrared spectroscopy that proteins with this type of alaninearomatic core are capable of forming coiled-coils (Figure 6). The conservation of the alanine-aromatic core between species suggests it to be an adaptation suited to ootheca fabrication or function in an as-yet-undetermined way. In some coiled-coils, Ala residues in the hydrophobic core and alternating core layer sizes are known to induce conformational distortions and regions of flexibility.48,51,52 Possibly, the alanine-aromatic core of mantis fibroins increases the flexibility of coiled-coils and of the material at the macroscopic level. We are currently investigating coiled-coil formation by mantis fibroins and their organization in native oothecae to better understand the role of their alanine-aromatic core.

■

ASSOCIATED CONTENT

S Supporting Information *

Figure S1 shows FTIR spectra obtained from ootheca made by each of the three study species. Hydrophobicity and side-chain lengths of amino acids according to heptad position are compared in Figure S2. Figure S3 shows FTIR spectra demonstrating artificial materials produced from recombinant Mantis Fibroin 1, Mantis Fibroin 2, and equimolar mixtures of both proteins, have similar secondary structure. Table S1A-C lists all cDNAs sequenced from collaterial gland libraries from the three species. Table S2 describes observed variations in mantis fibroin sequences attributed to allelic variation. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +61 2 6246 4236. Fax: +61 2 6246 4000. Email: Tara. [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Sri Sriskantha, Michelle Williams, and Peter Campbell for help with laboratory procedures and Holly Trueman and Paul Carr for reviewing the manuscript. We are grateful for comments from Dek Woolfson in regard to the coiled-coil components of these proteins. This project was supported by funding from the CSIRO and ANU.

■

■

CONCLUSIONS Mantis ootheca is an extensive extracorporeal material consisting of two coiled-coil proteins: Mantis Fibroin 1 and Mantis Fibroin 2. The two proteins have similar length, molecular architecture, and unusual coiled-coil core composition in all three species studied, implying that these are important functional features suited to the fabrication of mantis ootheca. However, the sequences of Mantis Fibroin 1 and 2 could not be convincingly aligned; therefore, it is unclear if their common features are the result of shared ancestry or convergent evolution. Mantis Fibroin 1 and Mantis Fibroin 2 are produced by each species studied at approximately equal levels, implying that each protein plays a critical role in the natural material. The relatively small size and diverse amino acid composition of mantis fibroins allows the production of full-length recombinant proteins. Coiled-coil protein solutions and solids could be produced from recombinant proteins, indicating that the two proteins were not required for protein folding. The mantis fibroin primary sequences had diverged considerably between species, indicating that the sequences

REFERENCES

(1) Monera, O. D.; Zhou, N. E.; Kay, C. M.; Hodges, R. S. J. Biol. Chem. 1993, 268 (26), 19218−19227. (2) Falvo, M. R.; Gorkun, O. V.; Lord, S. T. Biophys. Chem. 2010, 152, 15−20. (3) Rudra, J. S.; Tripathi, P. K.; Hildeman, D. A.; Jung, J. P.; Collier, J. H. Biomaterials 2010, 31, 8475−8483. (4) Woolfson, D. N. Adv. Protein Chem. 2005, 70, 79−112. (5) MacPhee, C. E.; Woolfson, D. N. Curr. Opin. Solid State Mater. Sci. 2004, 8, 141−149. (6) Kopecek, J. Biomaterials 2007, 28, 5185−5192. (7) Woolfson, D. N. Pept. Sci. 2010, 94, 118−127. (8) Sutherland, T. D.; Weisman, S; Walker, A. A.; Mudie, S. T. Biopolymers 2011, 97, 446−454. (9) Kojima, S.; Kuriki, Y.; Yoshida, T.; Yazaki, K.; Miura, K. Proc. Jpn. Acad., Ser. B 1997, 73, 7−11. (10) Dong, H.; Paramonov, S. E.; Hartgerink, J. D. J. Am. Chem. Soc. 2008, 130, 13691−13695. (11) Gribbon, C.; Channon, K. J.; Zhang, W.; Banwell, E. F.; Bromley, E. H. C.; Chaudhuri, J. B.; Oreffo, R. O. C.; Woolfson, D. N. Biochemistry 2008, 47, 10365−10371. (12) Pandya, M. J.; Spooner, G. M.; Sunde, M.; Thorpe, J. R.; Rodger, A.; Woolfson, D. N. Biochemistry 2000, 39, 8728−8734. 4271

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272

Biomacromolecules

Article

(51) Dunin-Horkawicz, S.; Lupas, A. N. J. Struct. Biol. 2010, 170, 226−235. (52) Gernert, K. M.; Surles, M. C.; Labean, T. H.; Richardson, J. S.; Richardson, D. C. Protein Sci. 1995, 4, 2252−2260.

(13) Potekhin, S. A.; Melnik, T. N.; Popov, V.; Lanina, N. F.; Vazina, A. A.; Rigler, P.; Verdini, A. S.; Corradin, G.; Kajava, A. V. Chem. Biol. 2001, 8, 1025−1032. (14) Wagner, D. E.; Phillips, C. L.; Ali, W. M.; Nybakken, G. E.; Crawford, E. D.; Schwab, A. D.; Smith, W. F.; Fairman, R. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 12656−12661. (15) Xu, C.; Kopecek, J. Pharm. Res. 2008, 25, 674−682. (16) Zimenkov, Y.; Conticello, V. P.; Guo, L.; Thiyagarajan, P. Tetrahedron 2004, 60, 7237−7246. (17) Smith, A. M.; Banwell, E. F.; Edwards, W. R.; Pandya, M. J.; Woolfson, D. N. Adv. Funct. Mater. 2006, 16, 1022−1030. (18) Ryadnov, M. G.; Woolfson, D. N. J. Am. Chem. Soc. 2005, 127, 12407−12415. (19) Banwell, E. F.; Abelardo, E. S.; Adams, D. J.; Birchall, M. A.; Corrigan, A.; Donald, A. M.; Kirkland, M.; Serpell, L. C.; Butler, M. F.; Woolfson, D. N. Nat. Mater. 2009, 8, 596−600. (20) Rockwood, D. N.; Preda, R. C.; Yucel, T.; Wang, X.; Lovett, M. L.; Kaplan, D. L. Nat. Protocols 2011, 6, 1612−1631. (21) Widhe, M.; Johansson, J.; Hedhammar, M.; Rising, A. Biopolymers 2012, 97, 468−478. (22) Sutherland, T. D.; Church, J. S.; Hu, X.; Huson, M. G.; Kaplan, D. L.; Weisman, S. PLoS ONE 2011, 6, e16489. (23) Weisman, S.; Haritos, V. S.; Church, J. S.; Huson, M. G.; Mudie, S. T.; Rodgers, A. J. W.; Dumsday, G. J.; Sutherland, T. D. Biomaterials 2011, 31, 2695−2700. (24) Kameda, T.; Kojima, K.; Togawa, E.; Sezutsu, H.; Zhang, Q.; Teramoto, H.; Tamada, Y. Biomacromolecules 2010, 11, 1009−1018. (25) Kameda, T.; Kojima, K.; Zhang, Q.; Sezutsu, H.; Teramoto, H.; Kuwana, Y.; Tamada, Y. Comp. Biochem. Physiol., Part B 2008, 151, 221−224. (26) Sutherland, T. D.; Campbell, P. M.; Weisman, S.; Trueman, H. E.; Sriskantha, A.; Wanjura, W. J.; Haritos, V. S. Genome Res. 2006, 16, 1414−1421. (27) Sutherland, T. D.; Weisman, S.; Trueman, H. E.; Sriskantha, A.; Trueman, J. W.; Haritos, V. S. Mol. Biol. Evol. 2007, 24, 2424−2432. (28) Rudall, K. M. Sci. Basis Med. 1956, 5, 217−30. (29) Bullough, P. A.; Tulloch, P. A. J. Mol. Biol. 1990, 215, 161−73. (30) Rattew, C. J. D. Ph.D. Thesis, University of Oxford, 1974. (31) Kramer, K. J.; Ong, J.; Law, J. H. Insect Biochem. 1973, 3, 297− 302. (32) Rudall, K. M. In Comparative Biochemistry; Florkin, M., Mason, H. S., Eds.; Academic Press: New York, 1962; p 397−433. (33) Kenchington, W.; Flower, N. E. J. Microsc. 1969, 89, 263−281. (34) Atkins, E. D. T. J. Mol. Biol. 1967, 24, 139−140. (35) Fahnestock, S. R.; Yao, Z.; Bedzyk, L. A. Rev. Mol. Biotechnol. 2000, 74, 105−119. (36) Polymeric Materials Encyclopedia; Salomone, J. C., Ed.; CRC Press: London, 1996. (37) Rice, P.; Longden, I.; Bleasby, A. Trends Genet. 2000, 16, 276− 277. (38) Hall, T. A. Nucleic Acids Symp. Ser. 1999, 41, 95−98. (39) Emanuelsson, O.; Brunak, S.; von Heijne, G.; Nielsen, H. Nat. Protocols 2007, 2, 953−971. (40) Thompson, J. D.; Higgins, D. G.; Gibson, T. J. Nucleic Acids Res. 1994, 22, 4673−4680. (41) Jones, D. T. J. Mol. Biol. 1999, 292, 195−202. (42) Delorenzi, M.; Speed, T. Bioinformatics 2002, 18, 617−625. (43) Eisenberg, D.; McLachlan, A. D. Nature 1986, 319, 199−203. (44) Heimburg, T.; Schunemann, J.; Weber, K.; Geisler, N. Biochemistry 1999, 38, 12727−12734. (45) Yoshimizu, H.; Ando, I. Macromolecules 1990, 23, 2908−2912. (46) Yoshimizu, H.; Mimura, H.; Ando, I. Macromolecules 1991, 24, 862−866. (47) Tuzi, S.; Sakamaki, S.; Ando, I. J. Mol. Struct. 1990, 221, 289− 297. (48) Brown, J. H. Protein Sci. 2010, 19, 1366−1375. (49) Yoder, N. C.; Kumar, K. J. Am. Chem. Soc. 2006, 128, 188−191. (50) Liu, J.; Yong, W.; Deng, Y.; Kallenbach, N. R.; Lu, M. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 16156−16161. 4272

dx.doi.org/10.1021/bm301570v | Biomacromolecules 2012, 13, 4264−4272