New Zealand Glowworm (Arachnocampa luminosa) Bioluminescence Is Produced by a Firefly-Like Luciferase but an Entirely New Luciferin

• A firefly-family luciferase paired with a completely novel luciferin in a fly. Arachnocampa luminosa, the New Zealand cave glowworm (a fungus gnat larva, Diptera: Keroplatidae), produces blue-green light (λmax 487 nm) using a 59 kDa enzyme that sits in the same ANL adenylating-enzyme superfamily as firefly luciferase but shares only ~30% sequence identity with P. pyralis and L. cruciata luciferases. The substrate, however, is not D-luciferin, it is a previously unknown molecule built from L-tyrosine and xanthurenic acid (XA, a kynurenine-pathway insect metabolite). The candidate luciferin “LRC” was isolated at 10 µg quantities, characterized by high-resolution MS, MS/MS fragmentation, deuterium-exchange MS, and ¹H NMR, and shown to produce intense fast-peak luminescence with crude luciferin-depleted glowworm luciferase. Diptera and Coleoptera diverged ~330 Mya with no known luminous lineages between, this is genuinely independent recruitment.
• Direct refutation of the Trowell et al. (2016) claim that Arachnocampa uses firefly D-luciferin. Watkins shows no cross-reaction in either direction (firefly luciferin + glowworm luciferase, or glowworm luciferin + firefly luciferase yield only background), and LC-MS of A. luminosa lysates detects no firefly D-luciferin at all. The earlier Trowell report of recombinant A. richardsae luciferase producing light with D-luciferin is reframed here as likely promiscuous activity, even non-luminous insect acyl-CoA enzymes have low-level luciferase activity with firefly luciferin or analogs (Mofford 2014). The cautionary lesson: ANL-superfamily proteins frequently show baseline luciferase activity, so “produces light with D-luciferin in a heterologous system” is a much weaker claim than “uses D-luciferin in vivo.”
• Tyrosine + xanthurenic acid as luciferin precursors is unprecedented. No other characterized bioluminescent system uses this chemistry. Tyr is a component of coelenterazine and the Siberian earthworm Fridericia heliota luciferin, and XA is a well-studied insect kynurenine-pathway metabolite (antioxidant, eye pigment precursor, chemiluminescent quinoline), but their combination into a luciferin appears unique to glowworms. The proposed candidate structure (Fig. 7C in the paper) is a Tyr-XA conjugate with an aldehyde at the XA C3 position; final structural confirmation is pending synthesis. The “slow-peak” vs “fast-peak” luminescence kinetics observed in different lysates are explained by a two-step model: free Tyr + XA require an unidentified luciferin synthetase to assemble LRC before the luciferase can act on it, while pre-formed LRC produces fast-peak light directly. This parallels the hispidin → 3-hydroxyhispidin precursor logic in fungi (Purtov 2015 / Kotlobay 2018).
• Multiple independent recruitments of acyl-CoA synthetase enzymes as luciferases across the tree of life. Watkins frames this as a major emerging pattern: bioluminescent beetles (fireflies, click beetles, railroad worms), the Japanese firefly squid Watasenia scintillans, and now the Arachnocampa glowworm all use ANL-superfamily / acyl-CoA synthetase enzymes to adenylate and oxidize structurally entirely different luciferins. The Siberian earthworm Fridericia heliota is also ATP-dependent and may be another instance. The authors' interpretation: the adenylation chemistry of acyl-CoA synthetases provides a uniquely flexible substrate for evolution of bioluminescent activity, because activating a carboxylate to an adenylate intermediate is a generic step that can be applied to many different luciferin scaffolds. This is convergent neofunctionalization at the enzyme-family level, distinct lineages independently coopting members of the same broad enzyme family for the same general chemistry on completely different substrates.
• Active-site divergence from beetle luciferase is significant despite the family-level relationship. ATP-binding motifs and the lysine residue critical to the firefly adenylation half-reaction are conserved in the glowworm luciferase, but the residues that bind D-luciferin in the firefly enzyme are not well conserved, and the lysine that drives the oxidation (light-producing) half-reaction in beetle luciferase is replaced by a methionine in Arachnocampa. So the ATP-handling machinery is shared but the substrate-binding pocket has been substantially rebuilt to accommodate the Tyr-XA scaffold. This is a useful negative result for anyone tempted to argue that “luciferase is luciferase,” the architecture is genuinely modular, with the adenylation half conserved and the oxidation half rebuilt around each substrate.
• Recombinant expression failed across multiple hosts. Bacterial, insect, and human cell expression of A. luminosa luciferase did not yield soluble protein at the time of publication. This is a recurring frustration with non-firefly luciferases, many do not fold well in E. coli, and is part of why firefly luc2 remains the dominant beetle-family reporter despite the glowworm system having an attractively blue-shifted emission for tissue penetration.

Bottom line for the project: This paper is the strongest single citation for the argument that ANL-superfamily / acyl-CoA synthetase enzymes are repeatedly recruited as luciferases across deep evolutionary distances, making firefly luc2 not just one option among many but a member of an enzyme family with a demonstrated propensity for evolving bioluminescent activity.