Single-cell pigment analysis of phototrophic and phyllosphere bacteria using simultaneous detection of Raman and autofluorescence spectra

Abstract

Microbes produce various types of pigments that are essential for their biological activities. Microbial pigments are important for humans because they are used in the food industry and medicine. The visualization and evaluation of the pigment diversity of microbial cells living in natural environments will contribute not only to the understanding of their ecophysiology but also to the screening of useful microbes. Here, we demonstrate the simultaneous, nondestructive detection of the resonance Raman and autofluorescence spectra of pigments in model purple phototrophic bacteria at the single-cell level. The single-cell Raman spectra measured using confocal laser Raman microspectroscopy with 632.8 nm excitation covered the wavenumber range of 660-3,022 cm^-1 (corresponding to 661-783 nm), in which the autofluorescence spectra from the pigments can be detected simultaneously as a baseline. The peak position of the resonance Raman spectra of the carotenoids in the cells provided information on the length of the polyene chain and structural characteristics, such as conjugated keto groups and terminal rings. By contrast, the extracted autofluorescence spectra of purple phototrophic bacteria differed in pattern depending on bacteriochlorophyll type (a or b), suggesting that their autofluorescence originates from bacteriochlorophyll-related molecules. In addition, we revealed the pigment diversity in microbial cells on the leaf surface and isolated pigmented bacteria that could contribute to the pigment diversity of the environmental sample. Our study shows that Raman and fluorescence microspectroscopy is a useful tool for finding novel pigmented microbes and uncovering yet unknown relationships between microbes and light.IMPORTANCETo understand the activities of microbes in natural environments, it is important to know the types of biomolecules they express in situ. In this study, we report a method using resonance Raman and autofluorescence signatures to detect and distinguish the types of carotenoid and bacteriochlorophyll pigments in intact, living cells. We have shown that this method can be used to estimate the expression status and pigment types in purple phototrophic bacteria and carotenoid-producing bacteria as well as the diversity of the pigments expressed by microbes on the leaf surface. Our method requires little pretreatment and can analyze pigments without destroying cells, making it a useful tool for visualizing phototrophic activity and searching for unidentified microbes.

Keywords: Methylobacterium; aerobic anoxygenic phototrophic bacteria; autofluorescence; bacteriochlorophyll b; carotenoids; purple phototrophic bacteria; resonance Raman spectroscopy.

Fig 1

Representative single-cell spectra of seven phototrophic bacteria, four nonphototrophic and carotenoid-producing bacteria, and a nonphototrophic and noncarotenoid-producing bacterium. The two graphs enclosed by gray-dotted lines represent the spectra of a cell of R. sphaeroides cultured under phototrophic and aerobic growth conditions, respectively. The single-cell spectra were measured every 30 s up to 90 s (0–30 s, black; 30–60 s, gray; and 60–90 s, light gray lines).

Fig 2

(A) Averaged Raman spectra of the seven phototrophic bacteria and four nonphototrophic and carotenoid-producing bacteria. (B) C = C stretching (ν₁) peak position vs. C–C stretching (ν₂) peak position of carotenoids detected in single-cell Raman spectra of the same species as in A. Of the 25 cells, cells with resonance Raman peaks of carotenoids are shown: 21 cells for B. viridis, 23 for R. sphaeroides, 19 for R. gelatinosus, 15 for R. litoralis, and 25 for each of the other species. (C) Box plot of the Raman peak intensity for each cell of the 11 bacterial species. The Raman intensity is relative to that of the strongest ν₂ peak of R. rubrum. A significance test was performed for each of the ν₁ and ν₂ peaks. Different letters indicate statistically significant differences (P < 0.05) among the 11 species using a nonparametric Kruskal–Wallis test followed by a post hoc Dunn–Holland–Wolfe multiple comparison test.

Fig 3

(A) Averaged autofluorescence spectra of the seven phototrophic bacteria. The spectra in the graph, from top to bottom, are those of four types of anaerobic phototrophic bacteria, two types of aerobic phototrophic bacteria that produce BChl a, and one type of anaerobic phototrophic bacteria that produce BChl b. (B) PCA score plot (PC1 vs. PC2) based on the autofluorescence spectra of cells from model phototrophic bacteria. (C) Box plot of autofluorescence intensity for each cell of the 11 bacterial species (for R. sphaeroides, both anaerobic and aerobic growth). Autofluorescence intensity is relative to that of the strongest autofluorescence of A. vinosum. Different letters indicate statistically significant differences (P < 0.05) among the 12 samples by a nonparametric Kruskal–Wallis test followed by a post hoc Dunn–Holland–Wolfe multiple comparison test.

Fig 4

(A) Box plot of autofluorescence intensity for each cell collected from clover leaves that showed carotenoid Raman peaks and for cells of B. subtilis as a nonphototroph. Autofluorescence intensity is relative to that of the strongest autofluorescence of a cell from clover leaves. (B) C = C stretching (ν₁) peak position vs. C–C stretching (ν₂) peak position of carotenoids in each cell collected from clover leaves. Of 85 cells, 23 cells (open circles) showed resonance Raman peaks of carotenoids without autofluorescence, whereas 62 cells (closed circles) showed carotenoid Raman peaks with autofluorescence. (C) The scatter plot depicts the ν₁ vs. ν₂ peak position of carotenoids of the model bacteria (Fig. 2B, gray symbols) overlaid with that of the clover sample (Fig. 4B, red circles).

Fig 5

(A) PCA score plot (PC1 vs. PC2) and X-means clustering based on the autofluorescence spectra of cells from clover leaves showing autofluorescence intensity above a threshold value. Different colors indicate clusters 1–3 (1, red; 2, yellow; and 3, blue), and cross marks indicate cluster centroids. (B) Autofluorescence spectra of individual cells in clusters 1–3.