Multifunctional warheads: Diversification of the toxin arsenal of centipedes via novel multidomain transcripts

Highlights

• First description of centipede-toxin transcripts that encode multiple peptide toxins

• First use of MALDI-imaging to examine toxin distribution/processing in venom glands

• Discovery of a rare class of predatory toxins that evolved under strong negative selection

• Discovery of a unique toxin expression strategy in venomous arthropods

Abstract

Arthropod toxins are almost invariably encoded by transcripts encoding prepropeptides that are posttranslationally processed to yield a single mature toxin. In striking contrast to this paradigm, we used a complementary transcriptomic, proteomic and MALDI-imaging approach to identify four classes of multidomain centipede-toxin transcripts that each encodes multiple mature toxins. These multifunctional warheads comprise either: (1) repeats of linear peptides; (2) linear peptides preceding cysteine-rich peptides; (3) cysteine-rich peptides preceding linear peptides; or (4) repeats of linear peptides preceding cysteine-rich peptides. MALDI imaging of centipede venom glands revealed that these peptides are posttranslationally liberated from the original gene product in the venom gland and not by proteases following venom secretion. These multidomain transcripts exhibit a remarkable conservation of coding sequences, in striking contrast to monodomain toxin transcripts from related centipede species, and we demonstrate that they represent a rare class of predatory toxins that have evolved under strong negative selection. We hypothesize that the peptide toxins liberated from multidomain precursors might have synergistic modes of action, thereby allowing negative selection to dominate as the toxins encoded by the same transcript become increasingly interdependent.

Biological significance

These results have direct implications for understanding the evolution of centipede venoms, and highlight the importance of taking a multidisciplinary approach for the investigation of novel venoms. The potential synergistic actions of the mature peptides are also of relevance to the growing biodiscovery efforts aimed at centipede venom. We also demonstrate the application of MALDI imaging in providing a greater understanding of toxin production in venom glands. This is the first MALDI imaging data of any venom gland.

Introduction

Proteins earmarked for secretion are typically produced as prepropeptides comprised of a signal peptide and one or two propeptide regions that are posttranslationally excised by endoproteases to yield a single mature protein product [1], [2], [3]. There are, however, a number of deviations from this scheme, including transcripts that lack propeptide-encoding regions and transcripts that encode multiple mature proteins [3]. For example, in both vertebrates and invertebrates, neuropeptides and hormones are commonly produced as multifunctional precursors containing a signal peptide and multiple copies of the neuropeptide or hormone separated by propeptide regions [4].

Multifunctional toxin transcripts, however, are exceedingly rare in most venomous taxa, with the reptilian clade Toxicofera being the notable exception. Various strategies leading to multifunctional toxin transcripts have evolved both convergently and divergently on several occasions within Toxicofera. These include the duplication events leading to precursors encoding tandem stretches of sarafotoxins in Atractaspis snakes [5], seven newly evolved bradykinin potentiating peptides in the propeptide region of the precursor encoding a C-type natriuretic peptide in the pit viper Bothrops jararaca [6], and multiple helokinestatin peptides in the propeptide region of the precursor encoding a B-type natriuretic peptide in Helodermatidae and Anguiidae lizard venoms [7], [8], [9]. The venom glands of coleoid cephalopods (cuttlefish, octopus and squid) also produce multifunctional transcripts encoding 3–4 pacifastin peptides that are posttranslationally liberated [10].

In striking contrast to coleoids and toxicoferans, invertebrate venomous animals such as marine cone snails, hymenopterans, sea anemones, scorpions, and spiders strictly adhere to the canonical one gene–one toxin strategy [11], [12]. The most widely studied venomous arthropods, namely spiders and scorpions, generate venom diversity via expression of numerous isoforms of each toxin type rather than via multiple posttranslational modifications of a single translated product [12], [13]. Their impressive toxin arsenal [14] appears to have evolved through classical gene duplication events followed by explosive diversification driven by positive selection [12], [15]. In contrast, transcripts encoding multiple mature toxins are extremely rare in arthropods, and have only been noted for latarcins, linear antimicrobial peptides found in the venom of the spider Lachesana tarabaevi [16].

Centipedes may be the oldest extant terrestrial venomous lineage, having arisen more than 400 million years ago (Mya) [17]. Reflecting this ancient divergence, the centipede venom apparatus as well as most centipede toxins described to date bear little resemblance to those of other arthropods [18], [19], [20]. However, the centipede toxin transcripts described to date conform to the arthropod paradigm of encoding a prepropeptide containing a single mature toxin domain. In striking contrast, we describe here four different types of multidomain transcripts from the venom gland of four species of scolopendrid centipede and use MALDI imaging to show that these multifunctional “warheads” are activated in the venom gland prior to venom expulsion.