A bacterial sulfonolipid triggers multicellular development in the closest living relatives of animals

Abstract

Bacterially-produced small molecules exert profound influences on animal health, morphogenesis, and evolution through poorly understood mechanisms. In one of the closest living relatives of animals, the choanoflagellate Salpingoeca rosetta, we find that rosette colony development is induced by the prey bacterium Algoriphagus machipongonensis and its close relatives in the Bacteroidetes phylum. Here we show that a rosette inducing factor (RIF-1) produced by A. machipongonensis belongs to the small class of sulfonolipids, obscure relatives of the better known sphingolipids that play important roles in signal transmission in plants, animals, and fungi. RIF-1 has extraordinary potency (femtomolar, or 10(-15) M) and S. rosetta can respond to it over a broad dynamic range-nine orders of magnitude. This study provides a prototypical example of bacterial sulfonolipids triggering eukaryotic morphogenesis and suggests molecular mechanisms through which bacteria may have contributed to the evolution of animals.DOI:http://dx.doi.org/10.7554/eLife.00013.001.

Keywords: Algoriphagus; Other; Salpingoeca rosetta; bacterial sulfonolipid; multicellular development.

Figure 1.. Rosette colony development in S. rosetta is regulated by A. machipongonensis.

(A) The original culture of S. rosetta, ATCC 50818, contains diverse co-isolated environmental bacteria and forms rosette colonies (arrowheads) rarely. (B) Treatment of ATCC50818 with a cocktail of antibiotics reduced the bacterial diversity and yielded an S. rosetta culture line, RCA, in which rosette colonies never formed. (Representative single cells indicated by arrows.) (C) Addition of A. machipongonensis to RCA cultures was sufficient to induce rosette development. Scale bar, 2 μm. DOI: http://dx.doi.org/10.7554/eLife.00013.003

Figure 1—figure supplement 1.. Frequency of rosette colonies in S. rosetta environmental isolate ATCC 50818, RCA with and without A. machipongonensis and a monoxenic line with A. machipongonensis feeder bacteria (Px1).

Altering bacterial diversity in S. rosetta cultures alters the frequency of rosette colonies. Data are the whisker-box plots of the frequency of colonial cells in ATCC 50818 and a monoxenic culture of S. rosetta fed only A. machipongonensis bacteria (Px1) for three experiments. DOI: http://dx.doi.org/10.7554/eLife.00013.004

Figure 2.. Diverse members of the Bacteroidetes phylum induce rosette colony development.

A maximum likelihood phylogeny inferred from 16S rDNA gene sequences reveals the evolutionary relationships among A. machipongonensis, other members of the Bacteroidetes phylum, and representative γ-proteobacteria (γ), α-proteobacteria (α), and Gram-positive (+) bacteria. All 15 members of the Algoriphagus genus (Table 1), as well as six other species in the Bacteroidetes phylum, were competent to induce colony development (filled squares). In contrast, no species outside of Bacteroidetes and most of the non-Algoriphagus bacteria tested failed to induce rosette colony development (open squares). Scale bar, 0.1 substitutions per nucleotide position. DOI: http://dx.doi.org/10.7554/eLife.00013.006

Figure 3.. RIF-1, a sulfonolipid that induces rosette colony development.

(A) Rosette colony development is induced by live A. machipongonensis and the sphingolipid-enriched lipid fraction (20 mg mL⁻¹), but not by fresh medium, A. machipongonensis LPS (10 mg mL⁻¹), PGN (50 mg mL⁻¹), or LPS+PGN. Shown are the whisker-box plots of the % colonial cells/total cells under each condition in three independent experiments. (B) The molecular structure of RIF-1 deduced from MS and 1D- and 2D-NMR data. The RIF-1 structure, 3,5-dihydroxy-2-(2-hydroxy-13-methyltetradecanamido)-15-methylhexadecane-1-sulfonic acid, has two parts: a base (shown in red) that defines the capnine, and a fatty acid (shown in black). Features that distinguish RIF-1 from other known capnoids are shown with colored arrows: the 2-hydroxy on the fatty acid (black) and the 5-hydroxy on the capnine base (red). DOI: http://dx.doi.org/10.7554/eLife.00013.008

Figure 3—figure supplement 1.. Separation of A. machipongonensis sphingolipids by thin layer chromatography (TLC).

Lipids enriched in sphingolipids were separated by TLC after visualization with ammonium molybdate in 10% H₂SO₄. Bands (1-12) as well as regions between bands (A-F) were tested for morphogenic activity. Region F possessed activity and was further purified. DOI: http://dx.doi.org/10.7554/eLife.00013.009

Figure 3—figure supplement 2.. MS/MS analysis of RIF-1.

A major fragment derived from m/z = 606 (M-H) in the MS/MS spectrum of RIF-1 corresponds to amino-sulfonic acid S1. HRMS m/z calcd for C₁₇H₃₆NO₅S (M-H): 366.23142. Found: 366.2310 (M-H)^-. DOI: http://dx.doi.org/10.7554/eLife.00013.010

Figure 3—figure supplement 3.. Key two-dimensional (2D) correlations of RIF-1: Observed COSY correlations.

Red double-head arrows show key ³J or ⁴J H-H correlations in the head regions (1 to 6 and 2′ to 3′) of the fatty acid and the capnine base and in the tail regions with geminal dimethyl groups (14 to 17 and 12′ to 15′). DOI: http://dx.doi.org/10.7554/eLife.00013.011

Figure 3—figure supplement 4.. Key two-dimensional (2D) correlations of RIF-1: Observed HMBC spin system.

Blue single-head arrows show key ²J or ³J H-C correlations in the head and tail regions of the fatty acid and capnine base. The correlations between C-1′ and H-2/N-H demonstrated that the fatty acid and capnine base are joined through an amide bond. DOI: http://dx.doi.org/10.7554/eLife.00013.012

Figure 3—figure supplement 5.. Key two-dimensional (2D) correlations of RIF-1: Observed TOCSY spin system.

Green bonds show two key spin systems in RIF-1 - HO-CH- in the fatty acid fragment and -CH₂-CH(NH)-CH(OH)-CH₂-CH(OH)-CH₂- in the capnine base fragment. DOI: http://dx.doi.org/10.7554/eLife.00013.013

Figure 3—figure supplement 6.. ¹H NMR spectrum of RIF-1.

The spectrum exhibits one NH (δ_H 8.21), three hydroxyl groups (δ_H 5.52, 5.20, and 4.31), five signals from 2.50-4.00 ppm (four methines connected to either nitrogen at δ_H-2 3.88 or oxygens at δ_H-2′ 3.80/δ_H-3 3.71-3.78/δ_H-5 3.53-3.61, and one methylene connected to sulfur at δ_H-1 2.56 & 3.01), twenty methylenes and four methyls (δ_H d, J = 6.6 Hz, 12H) in the high field region (δ_H 0.75-1.75 ppm). DOI: http://dx.doi.org/10.7554/eLife.00013.014

Figure 3—figure supplement 7.. gHMQC spectrum of RIF-1.

The ¹J H-C correlations demonstrate that 2 (δ_H 3.88, δ_C 50.89) is connected to a nitrogen; 2′, 3, and 5 (δ_H 3.80/δ_C 71.29, δ_H 3.71-3.78/δ_C 71.51, and δ_H 3.53-3.61/δ_C 70.20, respectively) are oxygenated; 1 (δ_H 3.01 and 2.56, δ_C 51.87) is adjacent to a sulfonic acid group; and all the other twenty methylenes and four methyls at high filed (δ_H 0.75-1.75/δ_C 22.00-42.00). DOI: http://dx.doi.org/10.7554/eLife.00013.015

Figure 3—figure supplement 8.. gCOSY spectrum of RIF-1.

Indicated are important H-H correlations between NH and H-2, 2′-OH and H-2′, 3-OH and H-3, and 5-OH and H-5. DOI: http://dx.doi.org/10.7554/eLife.00013.016

Figure 3—figure supplement 9.. Expanded dqfCOSY spectrum of RIF-1.

In panel A (δ_H 1.15-1.65 ppm/δ_H 3.52-3.82 ppm), H-3 (δ_H 3.73) shows correlation to H-4a (δ_H 1.30), H-5 (δ_H 3.57) to H-4a/H-4b (δ_H 1.30/1.47), and H-5 to H₂-6 (δ_H 1.20/1.29); H-2′ (∼δ_H 3.8) correlates to H₂-3′ (δ_H 1.45 and 1.54). Panel B (δ_H 2.4-4.0 ppm/δ_H 2.4-4.0 ppm) demonstrates the correlations between H₂-1 (δ_H 2.56/∼3.0) and H-2 (δ_H 3.88), and between H-2 and H-3. Panel C (δ_H 0.75-1.60 ppm/δ_H 0.75-1.60 ppm) exhibits correlations in the other methylenes and methyl groups. DOI: http://dx.doi.org/10.7554/eLife.00013.017

Figure 3—figure supplement 10.. Expanded dqfCOSY spectra of RIF-1.

A dqfCOSY spectrum was collected in order to get a clear connectivity in the oxygenated region (1-position to 6-position) in the capnoid base fragment. DOI: http://dx.doi.org/10.7554/eLife.00013.018

Figure 3—figure supplement 11.. gHMBC spectrum of RIF-1.

Indicated are important ²J or ³J H-C correlations between NH/H-2/H-2′ and C-1′ (δ_C 173.23), and between NH and C-2/C-3 (δ_C 50.89/71.51). Based on the MS/MS analysis, the fatty acid fragment must be 2-hydroxy-13-methyltetradecanoyl, and the capnine base fragment must be 2-NH-3,5-dihydroxy-15-methylhexadecane-1-sulfonate. Hence, the planar structure of RIF-1 is determined as shown in Fig. 3 and Fig. 3 – Figure Supplements 3-5. DOI: http://dx.doi.org/10.7554/eLife.00013.019

Figure 3—figure supplement 12.. Expanded gHMBC spectrum of RIF-1 (δ_H 0-4.00 ppm/δ_C 15.0-85.0 ppm).

Indicated are correlations between H-2 to C-4/C-1 (δ_C 41.36/51.87), between H-2′ and C-4′/C-3′ (δ_C 24.99/34.85), between H-1 and C-2/C-3 (δ_C 50.89/71.51), and between H-4 and C-5/C-3 (δ_C 70.20/71.51). DOI: http://dx.doi.org/10.7554/eLife.00013.020

Figure 3—figure supplement 13.. Expanded gHMBC spectrum of RIF-1 (δ_H 0.80-1.80 ppm/δ_C 20.0-40.0 ppm).

Key correlations are between H-16 and C-14/C-15/C-17 (δ_C 38.91/27.79/22.09), between H-17 and C-14/C-15/C-16 (δ_C 38.91/27.79/22.09), between H-14′ and C-12′/C-13′/C-15′ (δ_C 38.91/27.79/22.09), and between H-15′ and C-12′/C-13′/C-14′ (δ_C 38.91/27.79/22.09). DOI: http://dx.doi.org/10.7554/eLife.00013.021

Figure 3—figure supplement 14.. TOCSY spectrum of RIF-1.

The spectrum shows correlations from NH to H₂-1 and H-5 through H-2, H-3 and H₂-4, and from 3-OH to H-5 via H₂-4 and H₂-1 through H-3 and H-2. DOI: http://dx.doi.org/10.7554/eLife.00013.022

Figure 3—figure supplement 15.. Expanded TOCSY spectrum of RIF-1 (δ_H 0.50-4.25 ppm/δ_C 0.50-4.25 ppm).

Another important spin system is clearly demonstrated by the TOCSY correlations between H-2′ and H₂-3′/H-4′/H-5′ on the top left of the expanded spectrum. DOI: http://dx.doi.org/10.7554/eLife.00013.023

Figure 4.. Purified RIF-1 is active at plausible environmental concentrations.

RIF-1 concentrations ranging from 10⁻² to 10⁷ fM induce rosette colony development in RCA cultures. Frequency of rosette colony development was quantified in RCA cultures 2 days after treatment with a dilution series of purified RIF-1. Data are mean ± s.e. from three independent experiments. Line indicates non-linear regression of the RIF-1 activity profile. DOI: http://dx.doi.org/10.7554/eLife.00013.025

Figure 4—figure supplement 1.. Detection of purified RIF-1.

RIF-1 did not show absorbance at 210 nm (top panel), but the molecule (606 Da, M-H) was detected between 25 and 26 minutes (bottom panel). DOI: http://dx.doi.org/10.7554/eLife.00013.026

Figure 4—figure supplement 2.. Detection of RIF-1 in the conditioned medium of A. machipongonensis.

RIF-1 (bottom panel) was detected in the conditioned medium after the broth was concentrated 250 times. DOI: http://dx.doi.org/10.7554/eLife.00013.027

Figure 4—figure supplement 3.. Co-injection of concentrated conditioned medium with purified RIF-1.

The peak of the RIF-1 in the conditioned medium was enhanced after the sample was spiked with purified RIF-1 (bottom panel). DOI: http://dx.doi.org/10.7554/eLife.00013.028