Some marine bacteria stimulate the early life-cycle transition from larval to juvenile phases in corals (i.e., metamorphosis) [1,2,3]. These bacteria could be used to promote coral larval recruitment in the wild, cultivate corals for reseeding degraded reefs or the aquarium trade, or help test basic science questions in the laboratory. Single-species biofilms of Pseudoalteromonas bacteria, such as Pseudoalteromonas sp. strain PS5, have been shown to stimulate coral metamorphosis [2, 4, 5]. Moreover, the compound tetrabromopyrrole (TBP), extracted from Pseudoalteromonas bacteria or chemically synthesized, robustly promotes the metamorphosis of diverse coral species with and without attachment [2, 5,6,7] (Fig. 1A). However, an open question remains about whether live Pseudoalteromonas bacteria stimulate coral larval metamorphosis solely by producing TBP, or whether these Pseudoalteromonas strains produce yet unknown products that confer part or all of the stimulatory activity [6, 8].

Fig. 1: TBP biosynthesis in Pseudoalteromonas sp. PS5 induces metamorphosis of the stony coral Porites astreoides.
figure 1

A A model of TBP biosynthesis in Pseudoalteromonas sp. PS5 and its ability to induce coral metamorphosis. B Synteny of the bmp gene cluster in Pseudoalteromonas sp. PS5 (Genbank accession KR011923) created with EasyFig (V1.4.4) [23]. The 5’UTR for both bmp1 and bmp9 were cloned and fused to a Nanoluc (Nluc) reporter shown in blue. The bmp2 brominase is highlighted in red. C Growth curve of Pseudoalteromonas sp. PS5 wild type, ∆bmp2 and strains expressing plasmids bmp1-NLuc, bmp9-NLuc, and CP25-gfp. Optical density (OD) measurements were taken at 600 nm wavelength and graphed on a Log10 scale. Dotted lines correspond to the Gompertz non-linear regression fit for wild type and ∆bmp2. Plotted is the average of two biological replicates. D LC-MS/MS of Pseudoalteromonas sp. PS5 wild type and bmp2 knockout quantifying tetrabromopyrrole production (N = 3 extractions per treatment extracted from replicate 5 mL liquid cultures, one-tailed Mann–Whitney test, P = 0.05). Error bars show standard deviations. E Metamorphosis biofilm assays (%) with Porites astreoides larvae (10 larvae/ dish) in response to Pseudoalteromonas sp. PS5 wild type and ∆bmp2 strains. Controls include disks incubated in sterile Marine Broth (MB) media or filtered seawater (SW). N = 25 total dishes per treatment, including 3 experiments with either 5 or 10 replicates spanning two collection years. Combined morphogenesis significance is shown (Dunn’s multiple comparison of wild type vs ∆bmp2, Adjusted P < 0.0001). F Luciferase assays of bmp1 and bmp9 promoter activity under different modes of growth reported in relative luminescence units normalized by the optical density (RLU/OD600) and plotted on the Log10 scale. The 5’UTR of bmp1 and bmp9 were compared against the negative control background (Pseudoalteromonas sp. PS5 cells expressing a non-luminescent plasmid) as represented by the dotted line (Y = 60,422 RLU/OD600). Error bars show standard deviation of the mean. N = 4 biological replicates. G Strong and uniform expression of the genomic GFP tag in the Pseudoalteromonas transconjugants observed with fluorescence microscopy (top). The bottom panel is the wild type Pseudoalteromonas sp. PS5 for comparison. Scale bar is 5 µm.

Full size image

In Pseudoalteromonas sp. PS5 and phylogenetically related strains, TBP biosynthesis from L-proline is performed by three enzymes and a carrier protein encoded by the brominated marine pyrroles/phenols (bmp) gene cluster [9], specifically genes bmp1-4 [10] (Fig. 1B). We thus sought to create a TBP deletion mutant in Pseudoalteromonas sp. PS5 to query coral larval metamorphosis by genetically inactivating the bmp2 gene, which codes for the brominase that installs all four bromine atoms in TBP [10]. We determined that Pseudoalteromonas sp. PS5 is amenable to genetic manipulation by conjugation using a broad-host-range gfp reporter plasmid from the Marine Modification Kit (MMK) plasmid system [11]. We generated a Pseudoalteromonas sp. PS5 mutant with an in-frame deletion of the bmp2 gene (∆bmp2) following previously established methods for double-homologous recombination [12] (Fig. 1B). The modified strains exhibited a similar growth rate to wild type; however, the plasmid containing and bmp2 mutant strains persisted longer in stationary phase (Fig. 1C). The mutation of bmp2, as anticipated, resulted in the complete loss of TBP production in the ∆bmp2 strain in stark contrast to the wild-type culture that readily produces TBP (1468 ± 401 µM TBP, P = 0.05, one-tailed Mann–Whitney test) (Fig. 1D). These results confirm that the bmp2 gene is required for TBP production in Pseudoalteromonas sp. PS5 under the conditions tested.

We next tested whether bacteria lacking the bmp2 gene were able to stimulate the metamorphosis of the Caribbean coral, Porites astreoides, which has been shown previously to undergo metamorphosis in response to Pseudoalteromonas sp. PS5 and TBP [2]. When exposed to biofilms of wild type Pseudoalteromonas sp. PS5, we observed the metamorphosis of coral larvae consistent with previous findings [2, 4,5,6], both attached to the substrate (11.6% ± 7.2, Adjusted P = 0.0002, Dunn’s multiple comparisons test) and unattached (55% ± 13.1, Adjusted P < 0.0001, Dunn’s multiple comparisons test) (Fig. 1E, Supplementary Table S1). In contrast, total morphogenesis of coral larvae was reduced from 63.5% in wild type biofilm treatments to 3.5% in the ∆bmp2 biofilm treatment (Adjusted P < 0.0001, Dunn’s multiple comparisons test) (Fig. 1E, Supplementary Table S1). We observed attached (0.6% ± 2.9) and unattached (5.1% ± 10.4) metamorphosis in the ∆bmp2 treatment at rates comparable to the background in the media disk and the seawater disk (placed directly in the Petri dishes with filtered seawater) controls. Our results demonstrate that the effect of Pseudoalteromonas sp. PS5 on coral metamorphosis is primarily due to the production of TBP.

We then questioned whether different growth conditions affect the expression of the bmp genes in Pseudoalteromonas sp. PS5. To this end, we cloned the bmp1 and bmp9 promoters, fused them with a NanoLuciferase (NLuc) reporter gene, and conjugated the resulting plasmids into Pseudoalteromonas sp. PS5 (Fig. 1B) [11]. Luminescence was measured during each growth phase with the highest activity measured in late stationary and biofilm phases (Fig. 1F). The bmp1 promoter displayed a 203-fold increase in expression between early stationary and biofilm (Fig. 1F). The bmp9 promoter followed similar expression profiles, suggesting that the gene cluster may be co-regulated. We also tested broad-host-range promoters, which displayed at least a 457-fold activity range compared to assay background across all tested conditions (Supplementary Fig. S1). Our results suggest that the expression of TBP biosynthesis genes is strongest when bacteria exist in a slow-growth state when Pseudoalteromonas sp. PS5 might occur within surface-attached biofilms on the sea floor.

Pseudoalteromonas species are known to associate with marine eukaryotes and produce interesting antimicrobial metabolites [13], yet the study of these host-microbe and microbe-microbe interactions remains challenging due to the limited genetic tools for their manipulation. We therefore developed an integrative Tn10 transposon for use in Pseudoalteromonas sp. PS5, which is compatible with existing modular genetic toolkit parts [11, 14, 15] (Supplementary Fig. S2). With the Tn10 transposon, we generated Pseudoalteromonas sp. PS5 with gfp-tags integrated into the genome (Fig. 1G, Supplementary Fig. S2), which were confirmed by whole genome sequencing to identify genomic insertion loci (Supplementary Table S2).

In this work, the genetic engineering of a marine microbe enabled us to test a standing question about the role of TBP in coral metamorphosis. Our results represent the first characterization of a gene in a marine bacterium conveying a morphogenic effect in corals. Previous studies suggest that Pseudoalteromonas species may not be present in ecologically relevant concentrations that would stimulate coral metamorphosis [6]. However, TBP could be used as a molecule to elucidate mechanisms of coral morphogenesis. The strains and methodological advancements developed in this study could be helpful for dissecting how TBP stimulates metamorphosis with and without attachment in corals [16] and more broadly for studying TBP’s effects on eukaryotic cell physiology [17,18,19,20]. Furthermore, these approaches could be applied to inform responsible use of coral probiotic strain candidates encoding the bmp gene cluster[21]. Beyond TBP, our results demonstrate how bacterial genetics can help characterize genes and gene products from bacteria in the context of non-model marine microbial interactions, providing new techniques to interrogate the microbial ecology of Pseudoalteromonas spp. We hope this work will represent a step towards elucidating function in bacteria-coral interactions and will inform the use of bacteria for coral reef restoration [22].