Raman spectrometric discrimination of flexirubin pigments from two genera of Bacteroidetes

Authors

  • Jan Jehlička,

    Corresponding author
    1. Institute of Geochemistry, Mineralogy and Mineral Resources, Charles University, Prague, Czech Republic
    • Correspondence: Jan Jehlička, Institute of Geochemistry, Mineralogy and Mineral Resources, Charles University, Albertov 6, 12843 Prague, Czech Republic.

      Tel.: +420 221 95 1503;

      fax: +420 221 95 1429;

      e-mail: jehlicka@natur.cuni.cz

    Search for more papers by this author
  • Kateřina Osterrothová,

    1. Institute of Geochemistry, Mineralogy and Mineral Resources, Charles University, Prague, Czech Republic
    Search for more papers by this author
  • Aharon Oren,

    1. Department of Plant and Environmental Sciences, The Institute of Life Sciences, The Hebrew University of Jerusalem, Jerusalem, Israel
    Search for more papers by this author
  • Howell G. M. Edwards

    1. Centre for Astrobiology and Extremophiles Research, School of Life Sciences, University of Bradford, Bradford, UK
    Search for more papers by this author

Abstract

Flexirubins are specific polyene pigments produced by several genera of Bacteroidetes. Colonies and cell extracts of Flavobacterium johnsoniae and Flexibacter elegans have been investigated by Raman spectroscopy to show that this fast and non-destructive technique can be used to differentiate these pigments from carotenoids and to compare the flexirubin content of the two microorganisms. The presence or absence of certain distinguishing features in the CH combination band region at 2500–2750 cm−1 can assist in the discrimination between the two flexirubins investigated. Raman spectroscopy is thus a suitable tool not only to detect flexirubin pigments in bacterial cells, but also to further characterize the pigments present in members of the Bacteroidetes genera that are rich in flexirubins.

Introduction

Many members of the Bacteroidetes genera (Flexibacter, Cytophaga, Sporocytophaga and relatives) are coloured yellow-orange to pink-red due to the presence of pigments of the flexirubin group. Flexirubin pigments are aryl polyenes containing a polyenoic acid chromophore terminated by a p-hydroxyphenyl group and esterified with a dialkylated resorcinol (Fig. 1). The polyene chain is mainly biosynthesized from acetate and the conjugated phenyl as well as the three adjacent carbon atoms along the chain are derived from tyrosine (Achenbach et al., 1978; Reichenbach et al., 1981; Achenbach, 1987). The length of the polyene chain can vary between six and eight double bonds, and further variations occur in the presence or absence of a methyl- or chloro-substituent in the meta position of the conjugated phenyl groups and in the alkyl substituents of the esterified resorcinol. In some Bacteroidetes strains, flexirubins may be present together with carotenoid pigments. Carotenoid pigments are usually found in marine and halophilic members of the group, whereas flexirubin pigments are more frequent in clinical, freshwater or soil-colonizing representatives (Reichenbach et al., 1974; Achenbach et al., 1978; Reichenbach et al., 1981; Achenbach, 1987; Reichenbach, 1992; Bernardet & Bowman, 2006, 2010; Nakagawa, 2010).

Figure 1.

Structure of the flexirubin compounds from Flavobacterium johnsoniae (a) and Flexibacter elegans (b) (Oren, 2011).

We have explored the use of Raman spectroscopy as a specific, rapid and simple test for the presence of flexirubin pigments. Raman spectroscopy is a non-destructive technique for the provision of structural information and has been extensively used for the characterization of carotenoid pigments, which have an extremely strong Raman signal even in a non-resonant mode. Carotenoids have two strong Raman bands due to in-phase ν1 (C=C) near 1515 cm−1 and ν2 (C–C) stretching vibrations of the polyene chain near 1157 cm−1. The positions of both ν1 and ν2 bands in the spectra are dependent on the length of the polyene chain (number of conjugated double bonds; Gill et al., 1970; Merlin, 1985; Withnall et al., 2003). Raman spectroscopy has been extensively used for the characterization of carotenoid pigments in natural samples including simple 40-carbon carotenoids, more complex carotenoids such as salinixanthin of Salinibacter, and the C-50 bacterioruberin carotenoids of the Halobacteriaceae (Wynn-Williams & Edwards, 2000; Withnall et al., 2003; Edwards et al., 2005; Marshall et al., 2006, 2007; Buijtels et al., 2008; Fendrihan et al., 2009; Vítek et al., 2009, 2012; de Oliveira et al., 2010; Jehlička et al., 2013). We show here that flexirubin pigments can also be characterized by Raman spectroscopy, and that their spectra possess distinguishing features in the CH combination band region at 2500–2750 cm−1 that can assist in the differentiation between flexirubins and carotenoids, and also facilitate the discrimination between different flexirubins.

Materials and methods

Microbial strains, culture conditions, and sample preparation

Flavobacterium johnsoniae (DSM 2064T) and Flexibacter elegans (DSM 3317T) were grown in liquid media based on the composition of CY agar and Flexibacter medium (DSM media 67 and 344, respectively; http://www.dsmz.de). Cultures (200-mL portions in 500-mL Erlenmeyer flasks) were incubated with shaking (100 r.p.m.) at 30 °C. Cells from late exponential phase cultures were harvested by centrifugation (10 min, 10 000 g). Part of the collected material was dried by lyophilization; part was extracted with methanol/acetone (1 : 1, v/v), and the extracts were dried under reduced pressure in the dark.

Methods

Micro-Raman analyses of lyophilized powders or dried pigment extracts were performed on a multichannel Renishaw In Via Reflex spectrometer coupled with a Peltier-cooled CCD detector. Excitation was generally provided by a 514.5 nm Ar+ laser (1800 lines mm−1 grating). To achieve enhanced signal-to-noise ratios, 10–30 spectral scans were accumulated, each with a 20-s exposure time with laser power ranging between 30 and 100 mW. Spectra were recorded at a resolution of 2 cm−1 over the wavenumber range between 100 and 2000 and 2000 and 3500 cm−1. Other excitation wavelengths were used to complement the measurements at 514.5 nm, or where insufficient quality of the spectra were obtained with the Ar+ laser. A few Raman spectra were hence recorded also with 785 nm excitation (Renishaw diode laser; 1200 lines mm−1 grating) and 844 nm excitation (tunable Ar+ laser, 600 lines mm−1 grating; SpectraPhysics) of a Horiba Jobin Yvon microspectrometer. Spectra of reference pure carotenoids [astaxanthin, neoxanthin (ChromaDex), myxoxanthin (LGC Standards), zeaxanthin, lutein, fucoxanthin (Sigma-Aldrich), cantaxanthin and echinenone (CaroteNature)] were recorded under identical conditions. Polystyrene, sulfur, diamond, ε-caprolactone, acetonitrile/toluene (50/50 v/v), cyclohexane, stearic acid and glycerol were used to check the calibration of the instrument. Raman spectra were exported into the Galactic *.SPC format. Spectra were compared using grams ai (Version 8.0; Thermo Electron Corp., Waltham, MA). Raman spectra were not subjected to any data manipulation or processing techniques and are reported as collected.

Results and discussion

First- and second-order Raman spectra of F. johnsoniae and F. elegans are shown in Fig. 2, and the Raman bands observed are listed in Tables 1 and 2. For comparison, Raman shifts of pure crystalline carotenoids (astaxanthin, neoxanthin, myxoxanthophyll, zeaxanthin, lutein, fucoxanthin, canthaxanthin and echinenone), recorded using the same experimental setup, are listed in Table 3.

Table 1. Raman bands observed in Flavobacterium johnsoniae for three excitation wavelengths
Lyophilized materialExtractsTentative assignments
844 nm785 nm514.5 nm785 nm514.5 nm
  1. m, medium; s, strong; sh, shoulder; w, weak.

3060 m 3057 s 3054 sCCH aromatic ring stretching
2668 2656 s 2656 sCombination band terminal CH3–C aliphatic chain
2319 w    
2267 m 2262 s 2268 s
1529 s1527 s1529 s1526 s1528 sν C=C carotenoid
1208 sh    CCC chain stretching
1189 w  1186 w1190 wCH” deformation
1158 m1156 s1154 s1153 s1153 sν C–C carotenoid
1135 m1133 s1133 s1132 s1134 sCC stretching saturated aliphatic chain
1007 m1004 m1004 m1003 m1004 sδ C=CH conjugated chain
Table 2. Raman bands observed in Flexibacter elegans for two excitation wavelengths
Lyophilized materialExtractsTentative assignments
785 nm514.5 nm785 nm514.5 nm
  1. br, broad; m, medium; s, strong; sh, shoulder; w, weak; vw, very weak.

 3025 m  C=CH stretching; CCH aromatic stretching
 2663 s 2660 sCombination band terminal CH3–C aliphatic chain
 2516 s 2522 sCombination band CH3 aromatic ring
 2306 m 2304 m
 2157 m 2158 m
 1581 w 1579 wCOH bending
1512 s1515 s1513 s1516 sν C=C carotenoid
 1449 w, br 1448 w, brCH2 deformation
1351 vw13531352 vw1354 vwCH2 deformation
1284 m1285 w, br1280 m1283 w, br 
 1211 1210CCC chain stretching
1188 w1187 w1187 w1191 wCH” deformation
1155 s1155 s1154 s1155 sν C–C carotenoid
1138 1138 CC stretching saturated aliphatic chain
1005 m1004 s1061 vw δ C=CH conjugated chain
 996 m, sh1003 m1005 sCC symmetric stretching aromatic ring
979 979 CH3 rocking
 960 m953960 vwCH3 rocking
 875 m872 vw COC glycidyl linkage stretching
Table 3. Raman bands observed for reference carotenoids. All measurements were carried out with 514.5 nm excitation, except for neoxanthin and myxoxathophyll (457 nm excitation)
AstaxanthinNeoxanthinMyxoxanthophyllZeaxanthinLuteinFucoxanthinCanthaxanthinEchinenone
  1. br, broad; m, medium; s, strong; sh, shoulder; w, weak; vw, very weak.

  2. Main carotenoid Raman bands in bold.

   878 vw872 vw   
   961 w966 w965 w, br  
1006 m 1000 m 1000 m 1008 m 1007 m 1013 w, br 1009 m 1006 m
1156 s 1154 s 1153 s 1159 s 1158 s 1162 m 1158 s 1156 s
1191 w, sh 1194 w, sh1192 w, sh1191 wsh1187 m, sh1192 w, sh1193 w, sh
   1215 vw, sh1212 wsh  1212 vw
1276 w 1278 w (532)1269 w1272 w1267 vw1274 w 
1447 w       
1513 s 1534 vs 1512 s 1521 s 1524 s 1529 s 1515 s 1512 s
   2166 m2165 m  2165 w, br
   2315 m2316 m   
   2346 w, sh2345 wsh  2309 w, br
   2477 w 2492 w, br  
   2525 m2529 m2547 w, br 2516 w, br
2671 m, br  2675 s2679 s2684 m, br2673 m, br2664 m, br
       2782 w
3032 m, br  3034 m3039 m3058 m, br3034 m, br3022 m, br
Figure 2.

First-order (left panels) and second-order (right panels) Raman spectra of Flavobacterium johnsoniae (a) and Flexibacter elegans (b).

The chemical structure of both pigments is based on an eight-unit C=C conjugated unsubstituted skeletal backbone terminating in a para-substituted aromatic phenolic ring, the other end of the C=C framework being linked through a ketonic ether linkage to a meta-substituted phenolic aromatic ring which contains one or more aliphatic residues (Fig. 1). Therefore we can expect a number of common features to dominate the Raman spectra: C=C and C–C stretching bands from the conjugated unsaturated chain backbone (similar to an unsubstituted carotenoid), aromatic CH and CC stretching and bending bands, CO stretches from the aromatic phenolic groups, aliphatic chain CCC stretching and bending modes and modes associated with the CCO.OC ketonic ether (aromatic ester type) linkage. Expected vibrational spectroscopic differences in the chemical moieties for the flexirubins from F. johnsoniae and F. elegans include:

  1. Aromatic C–CH3 stretches, ortho- and meta-substituted, in a phenolic ring occurring at about 900 cm−1 and the associated methyl deformation modes around 1400 cm−1. Another diagnostic difference is the observation of methyl rocking vibrations near 950 cm−1.
  2. Although the CH2 chain lengths for the aliphatic hydrocarbon side chains in both flexirubins differ, diagnostically it will be difficult to discriminate between them on this basis because the individual CH2 asymmetric and symmetric stretching and deformation modes will be convolved into broad band envelopes in the ranges 2830–2970 and 1350–1450 cm−1.
  3. A clear observable difference should be a feature assignable to the terminal CC(CH3)2 group in the F. johnsoniae flexirubin, which is not present in the F. elegans flexirubin, and for this we might expect stretching bands near 2850–2750 cm−1, deformation bands near 1350–1450 cm−1, and CCC bending modes near 500 cm−1. We would, however, expect some accidental degeneracies with the methyl group vibrations in the other flexirubin. All expected features, as listed above, were observed in the spectra recorded in lyophilized cell powders and in pigment extracts (Tables 1 and 2). The resonances were significantly different from those of reference carotenoids (Table 3).

The differences in the molecular structures of the F. johnsoniae and F. elegans flexirubins can be summarized as follows: the latter contains a methyl group substituent ortho to the phenolic hydroxyl group in one aromatic ring and meta to the hydroxyl group in the other aromatic ring. The F. elegans molecule also has a C12H26 side chain ortho to the hydroxyl group and the aromatic ether linkage, whereas the analogue in F. johnsoniae has a C11H23 aliphatic chain in the same position terminating in a C(CH3)2 functionality and a further C5H11 substituent meta to the hydroxyl group and an ether linkage in place of the simple ethyl group in the F. elegans flexirubin. A key feature which will assist materially in the discrimination between these flexirubins based on Raman spectra is the observation of weaker features in the CH combination band region 2500–2750 cm−1. Thus far the use of this feature has been much under-appreciated in Raman diagnostics of carotenoids and structurally related pigments because other vibrational features generally are present that can assist in the identification. However, as described above, the structural differences between the two flexirubins, based on the presence of two terminal methyl groups attached separately to the aromatic rings in the F. elegans flexirubin compared with the two methyl groups in the terminal aliphatic C(CH3)2 moiety of the F. johnsoniae pigment, would lead to identifiably different Raman bands in the combination band region involving fundamental CH and CC modes in combination (Lawson et al., 1995). Hence we can assign the bands at 2663/2660 and 2522/2516 cm−1 to the combination bands caused by CC and CH stretching arising from the CH3 terminal to the aliphatic chain and aromatic ring methyl groups, respectively, in the F. elegans pigment, whereas the analogue is found in the F. johnsoniae at 2658/2656 cm−1 only, reflecting the absence of the aromatic ring methyl groups found in the former. The CH stretching bands expected for the terminal methyl groups and aliphatic chain would be expected around 2970–2850 cm−1, but these were not differentiated here. Multiple assignments for the CH2 deformations cited in the table reflect the different molecular motions possible for these units in the aliphatic chains, namely scissoring, rocking, twisting and wagging, which encompass a broad range extending from 1470 to 950 cm−1 and which in molecules of this complexity will almost certainly involve mode coupling with CCC, COC and with aromatic ring stretching and deformation modes in their descriptors. It must be taken into account that in biological materials the possible binding of carotenoids to other cell components may lead to minor shifts of the positions of Raman bands (de Oliveira et al., 2010); thus, caution is recommended when interpreting minor band shifts. However, such shifts were negligible in our case, as the band positions for the pigment extracts were nearly identical to those measured in lyophilized cell powders (Tables 1 and 2).

We here assign Raman bands to flexirubin pigments for the first time. As for carotenoids, Raman resonance signals are very strong, so that the analysis does not depend on prior extraction and purification of the pigments. Like all Raman spectroscopy-based methods, the analysis is very fast. A simple test described in the literature for the presence of flexirubin pigments is based on a characteristic colour reaction when treated with 20% KOH. Colonies of bacteria containing flexirubrin-type pigments exhibit an immediate colour shift from yellow to orange or red, purple or brown, and revert to their initial colour when flooded with an acidic solution. However, such a colour change is not absolutely specific for flexirubin pigments (Bernardet et al., 2002; Oren, 2011). Raman spectroscopy can therefore be considered to be a more reliable test, with more informative results.

We have thus shown that Raman spectroscopy is a suitable tool not only for the sensitive detection of flexirubin pigments in bacterial cells, but also for the further characterization of the pigment or pigments present in cultures of members of the Bacteroidetes rich in flexirubins.

Acknowledgements

We thank Lily Mana for her assistance in culturing the microorganisms and Ota Frank for some of the measurements. This work was supported by grant no. P210/10/0467 from the Grant Agency of the Czech Republic and by institutional support MSM0021620855 from the Ministry of Education of the Czech Republic (to J.J.).

Ancillary