UMR-CNRS 7138, Systématique-Adaptation-Evolution, Equipe Symbiose, Université des Antilles et de la Guyane, UFR des Sciences Exactes et Naturelles, Département de Biologie. B.P. 592. 97159 Pointe-à-Pitre Cedex, Guadeloupe, France
UMR-CNRS 7138, Systématique-Adaptation-Evolution, Equipe Symbiose, Université des Antilles et de la Guyane, UFR des Sciences Exactes et Naturelles, Département de Biologie. B.P. 592. 97159 Pointe-à-Pitre Cedex, Guadeloupe, France
Background information. Marine nematodes belonging to the Stilbonematidae (Desmodoridae) family are described as living in obligatory association with sulfur-oxidizing chemoautotrophic ectosymbionts. The symbiotic bacteria carrying out this chemosynthesis should contain elemental sulfur in periplasmic granules as sulfur granules of chemoautotrophic endosymbionts described in various marine invertebrates.
Results. Based on TEM (transmission electron microscopy) analyses, extracellular bacteria surrounding Eubostrichus dianae possess these spherical periplasmic granules. Few investigative techniques can be used to identify elemental sulfur, S8, such as EDXS (energy dispersive X-ray spectroscopy) and EELS (electron energy loss spectroscopy), which are associated with cryo-fixation of the sample to avoid sulfur loss. These techniques are time consuming, expensive and require technical skills. Raman microspectrometry applied to the analysis of E. dianae allowed us to detect elemental sulfur, S8, and confirmed the location of these sulfur clusters in the bacterial coat. In the same way, Raman spectrometry was positively applied to the endosymbiotic bivalve Codakia orbicularis, suggesting that this technique can be used to characterize sulfur in ecto- as well as in endo-symbiotic sulfur-oxidizing bacteria.
Conclusions. As Raman spectrometry can be used on living organisms (without preliminary fixation) without sample damage and preserving the molecular structure of the sulfur (denatured during chemical fixation), it represents a very well-adapted investigative tool for biologists. This technique therefore permits us to detect quickly and easily (in a few seconds and on entire living animals) the presence of sulfur compounds in the symbiotic nematode.
Several marine organisms living in sulfidic environments are associated with symbiotic bacteria. In the benthic deep-sea environment of hydrothermal vents, cold seeps and whale bones, many associations between marine invertebrates and proteobacteria have been described, such as in arthropods (Bresiliidae) (Polz and Cavanaugh, 1995; Zbinden et al., 2004), polychaetous annelids (Alvinellidae) (Desbruyères et al., 1983; Cary and Stein, 1998) or in bivalve molluscs (Mytilidae, Vesicomyidae) (Felbeck et al., 1981; Cavanaugh, 1983; Deming et al., 1997) etc. Similar relationships have also been described in tropical coastal environments in bivalvia (Lucinidae) (Berg and Atalo, 1984; Reid, 1990) and in nematoda (Stilbonematidae) (Ott and Novak, 1989; Ott et al., 1991). These thiotrophic organisms depend on the chemoautotrophic metabolism of reduced sulfur compounds and form symbiotic relationships with sulfur-oxidizing bacteria (Vetter, 1991).
The nature of invertebrate bacterial symbionts living in a low-sulfidic environment is classically determined using analytical electron microscopy [SEM (scanning electron microscopy), TEM (transmission electron microscopy) and EDXS (energy dispersive X-ray spectroscopy)], biochemical analyses, enzymatic studies and phylogeny (Cavanaugh, 1985; Polz et al., 1992; Bauer-Nebelsick et al., 1996a, 1996b; Rinke et al., 2006). The studies carried out on stilbonematids supposed that bacterial symbionts are thio-autotrophic γ-proteobacteria (Ott et al., 1991; Polz et al., 1992; Ott et al., 2004); however, only two studies were reported using 16S rDNA (ribosomal DNA) sequence analysis. The first results obtained on the nematode Laxus sp. permitted the design of a specific probe from a single bacterial sequence (Polz et al., 1994). The second phylogenetic analyses on the nematode Eubostrichus dianae have not really been conclusive (Polz et al., 1999). Owing to the microscopic size of nematodes, all techniques used for the DNA extraction of the symbionts require a large quantity of worms (100–200 individuals per assay). Moreover, filamentous bacteria that densely covered the entire surfaces of the nematodes were not detected by PCR-based analysis. So, specific probes and/or primers of these ectosymbiotic bacteria cannot be designed. Nevertheless, ultrastructural studies (TEM) of these nematodes indicate that such ectosymbiotic bacteria contain spherical cytoplasmic inclusions (Polz et al., 1992). These electron-lucent inclusions resemble sulfur globules found in the chemoautotrophic symbionts described in some bivalves (Vetter, 1985) and oligochaetes (Giere et al., 1988), as well as free-living sulfur-bacteria such as Beggiatoa (Lawry et al., 1981). These data suggest that such bacteria belong to the sulfur-oxidizing group.
Additional techniques focusing on the functional genes [APS (5′ adenylylsulfate) reductase] and the detection of elemental sulfur by high-pressure freezing of organisms, followed by a freeze-substitution and cryo-embedding before a cryo-EFTEM (energy-filterered TEM) microanalysis, have recently been published (Lechaire et al., 2006, 2008). The main problem for the detection of sulfur is that the solvents, such as ethanol, used during dehydration of samples before epoxy-resin embedding dissolve sulfur compounds (Truchet et al., 1998; Pasteris et al., 2001). Moreover, the techniques used for this detection require a high technical skill, heavy sample preparation, and are consequently time consuming and expensive.
White et al. (2006) have shown the capabilities of Raman spectrometry to investigate sulfur S8 in biological materials by in situ analysis in deep ocean. The present study is concerned with the detection, identification and location of elemental sulfur in freshly collected biological samples by means of Raman microspectrometry. The main advantages of this visible light spectrometry are that it can be quickly applied to biological samples in their original environment (i.e. in seawater) without any specific preparation. The spectra can be acquired in a short time (few seconds) on small areas without sample damage.
The technique is applied to the ectosymbiotic nematode E. dianae (Hopper and Cefalu, 1973) for which no data are available concerning sulfur contents and to the endosymbiotic bivalve Codakia orbicularis (Linné, 1758) known to harbour sulfur-oxidizing gill-endosymbionts (Berg and Alatalo, 1984; Frenkiel and Mouëza, 1995), containing elemental sulfur granules as recently determined by EELS (electron energy loss spectroscopy) (Lechaire et al., 2008). Unlike the study of Pasteris et al. (2001) our present study focuses directly on entire organs or small individuals (meiofauna).
Results and Discussion
Characteristics of the nematode
The nematode E. dianae presents extremely long filaments which form a coat around the body of the worm (Figures 1a and 1b). The ectosymbiotic bacteria comprising this coat have a length of 80–100 μm and a diameter of 0.5–1 μm. We can observe on Figure 1(a) the white appearance of the bacterial coat in optical microscopy and in Figure 1(b) the uniform shape of bacteria arranged regularly on the cuticle of the nematode.
Raman analysis applied to a living nematode
In order to detect and localize sulfur on this ectosymbiotic nematode 30 Raman spectra were recorded across the living nematode. The laser probe diameter was 10 μm and the distance between two successive analyses was 5.6 μm. The white circles drawn on the SEM micrograph (Figure 1b) illustrate the size and location of the probe used to acquire the Raman profiles on the living animal.
Figure 1(c) shows a selection of three representative Raman spectra recorded in seawater (near the nematode), on the bacterial coat and in the middle part of the nematode (including the probed volume of the nematode body and twice the bacterial coat) respectively. The Raman spectra obtained on the bacterial coat and the midbody of the nematode present three Raman bands (at 160 cm−1, 225 cm−1 and 480 cm−1) characteristic of the presence of sulfur S8 (Poborchii, 1996). These bands are not present in the spectra of seawater (Figure 1c).
As already mentioned in previous work (Long, 1977), the intensity of a Raman scattering band characteristic of a compound is linearly related to its amount contained in the analysed volume. This fact is used to locate S8 in E. dianae. The diagram shown in Figure 1(d) represents the variation of the intensity of the S8 480 cm−1 Raman band as a function of the position of the laser probe on the sample (equivalent diagrams are obtained using the intensity of the two other peaks at 160 cm−1 and 225 cm−1).
The zero intensity points on this diagram (no sulfur) correspond to the analyses outside of the sample. Sulfur is detected over a distance of 100 μm corresponding to the diameter of the body of the nematode (40 μm) surrounded by its bacterial coat (thickness 30 μm). The intensity profile across the sample reveals two maxima separated by a valley corresponding to the centre of the nematode.
Spatial interpretation of the Raman results
In order to interpret this diagram, the nematode and its bacterial coat are schematically drawn in Figure 1(e). The body of the nematode is represented by a 40 μm diameter cylinder surrounded by the bacterial coat of 30 μm thickness. The laser probe is represented by a vertical cylinder crossing the sample. For simplification, sulfur is considered to be homogeneously dispersed at the probe scale level, at a concentration cb in the nematode body and cc in the bacterial coat.
The Raman scattered intensity is proportional to the exciting line intensity and to the amount of sulfur contained in the probed volume, i.e. the intersection volume between the cylinder representing the probe and the schematized sample multiplied by the local sulfur concentration (see the section ‘Quantitative interpretation of the spectra’ below). In the case of a perfectly transparent sample, the intensity of the incident exciting line is considered constant across the sample and the Raman scattering intensity is just proportional to the probed volume multiplied by the local concentration. In the case of an absorbing sample the results must be corrected for absorption phenomena. The local exciting line intensity and Raman scattering damping must be calculated taking into account the absorbing properties of the medium. The experimental and mathematical developments for this purpose are reported below in the section ‘Influence of the light absorption by the various parts of the sample on the Raman scattering intensity’.
Figure 1(f) presents theoretical Raman intensity profiles for the two following cases: cb=cc=1 [a.u. (arbitrary unit)] and cb=0 & cc=1 (a.u.). The theoretical profiles are calculated assuming a perfectly transparent sample and considering an absorbing sample using the mathematical expression and experimental absorption coefficients presented in the paragraph ‘Mathematical expression of the Raman scattered light intensity for an idealized ectosymbiotic nematode’.
The main effects of the absorption properties of the various parts of the sample (bacterial coat, nematode body and digestive tube) is the important reduction (∼50%) of the Raman scattered intensity. The difference between the two theoretical corrected profiles [cb=cc=1 (a.u.) and cb=0 & cc=1 (a.u.)] is the relative intensity changes (Δ) between Raman intensity maxima IM and minima Im:
For cb=cc=1 (a.u.), Δ = 0.27, and for cb=0 & cc=1 (a.u.), Δ = 0.38.
This last value agrees well with the experimental variation recorded on the sample (Δ=0.41) and strongly supports the absence of sulfur in the body of the nematode and its digestive tube.
This conclusion is confirmed by the Raman spectra recorded on a nematode where the bacterial coat was accidentally eliminated during sample manipulation. In that case, the spectra do not exhibit any characteristic line of sulfur, demonstrating that the body of the nematode is effectively sulfur free (results not shown).
The non-symmetrical shape of the experimental intensity profile can be attributed to the irregular distribution of the bacteria around the body of the nematode, as pointed out on the SEM micrograph (Figure 1b).
Raman and EELS investigations on embedded nematodes
Thus Raman microspectrometry analyses carried out on the nematodes E. dianae allowed us to detect the elemental sulfur S8 in living samples and to localize this compound in the bacterial coat. Owing to the size of the probe compared with the size of the bacteria, it was not possible to show whether this elemental sulfur was located inside (i.e. in the cytoplasm or within granules inside the periplasmic space) or outside the bacteria as previously described in symbiotic bivalves with cryo-EFTEM microanalysis+EELS (Lechaire et al., 2006, 2008).
In order to emphasize the advantages of Raman microspectrometry applied to living nematodes, compared with classical ATEM (analytical TEM), Raman microspectrometry is applied to a semi-thin section of classically prepared samples in order to demonstrate the loss of sulfur induced by the sample preparation process.
The light and transmission electron micrographs presented in Figures 2(a) and 2(b) were obtained on cross-sections of nematodes embedded in LR White resin. Semi-thin and thin sections for Raman and ATEM investigations were both obtained from the same block.
A Raman scan analysis was performed on the semi-thin section of E. dianae. Twenty spectra were recorded across the LR White resin, the bacterial coat and the body of the nematode (Figure 2a). The study of the twenty spectra obtained during the scan demonstrated the lack of the three characteristic peaks of sulfur (Figure 2d), pointing out the absence of sulfur S8 in the sample.
The same result was observed on samples analysed by EELS. On the ultra-thin section, no sulfur was detected in the body of the nematode, in the bacterial coat or in the granules included in each bacterial symbiont.
As expected, neither EELS (Figure 2c) nor Raman spectroscopy (Figure 2d) analyses carried out on various parts of the sample allowed us to detect sulfur. This clearly confirms that sample preparation for ATEM investigation is responsible for the loss of sulfur. Sulfur S8 is known to be soluble in ethanol, a solvent commonly used in cTEM (conventional TEM) after chemical fixation, during ethanol dehydration and epoxy-resin embedding of biological samples (Vetter, 1985; Truchet et al., 1998; Pasteris et al., 2001). Consequently, the small electron-lucent granules within the bacterial symbiont usually appear empty in ultra-thin sections.
The observations show that these techniques are inefficient on embedded organisms after chemical fixation; however, EELS on cryo-fixed samples allows precise location of compounds (Lechaire et al., 2006; 2008). Nevertheless, Raman microspectrometry is the easiest technique which provides a positive result in a few seconds from an entire animal and without heavy sample preparation, permitting an easy and fast screening of sulfur in biological samples.
Raman analysis on endosymbiotic bivalves
In order to confirm the efficiency of Raman microspectrometry for detection of elemental sulfur in any kind of sulfur-oxidizing symbiotic organisms, we applied this technique to an endosymbiotic model. We used the tropical bivalve C. orbicularis, which lives in the same low sulfidic environment as E. dianae but harbours intracellular thiotrophic bacteria inside specific cells of the gill tissue (Frenkiel and Mouëza, 1995). In the eulamellibranch bivalve C. orbicularis, the gills are thick and cover the visceral mass almost completely (Reid, 1990). As previously described by Frenkiel and Mouëza (1995), bacteriocytes represent one of the major cell types of the gill filament (Figure 3b) of C. orbicularis. The bacteriocytes, mostly located in the one-third of the lateral zone, possess a cytoplasm filled with intracellular bacteria characterized by numerous periplasmic empty vesicles (Figure 3c). This host cell is characterized by an apical pole bearing short microvilli in contact with the circulating seawater and few mitochondria as host organelles. Bacteria are usually individually enclosed and present the double membrane typical of Gram-negative bacteria (Figure 3c). The bacterial cytoplasm usually contains essentially DNA, ribosomes and empty clear granules located in the periplasmic space (Figure 3c) which are considered as sulfur granules according to their appearance with cTEM.
The inset in Figure 3(a) shows the location of the probe used to acquire the sulfur concentration profile. The two spectra in Figure 3(d) correspond to analyses in water and in the gill of C. orbicularis. Sulfur is clearly detected in this tissue, whereas no sulfur is detected in the surrounding water. The full intensity profile of the sulfur 225 cm−1 characteristic Raman line as a function of the probe location is presented in Figure 3(e). This obviously demonstrates the presence of sulfur in the gills of C. orbicularis which was previously located in the periplasm of the sulfur-oxidizing bacterial endosymbionts by Lechaire et al. (2008).
Detection and characterization of sulfur in thioautotrophic bacteria is of interest in the understanding of symbiotic relationships in marine invertebrates colonizing sulfidic environments.
Further investigations will focus on the use of such techniques to assume potential sulfur-oxidizing bacterial symbiotic interactions in organisms belonging to the meiofauna not yet described as symbiotic.
In the present study we demonstrated that Raman microspectrometry is a powerful tool for such a purpose. It allowed us to point out the presence of elemental sulfur in the biological samples, to identify its structure as the S8 species and to locate it in the samples with an actual spatial resolution of 10 μm.
It is hoped that this resolution will be improved down to 2 μm in the next few months. This ultimate spatial resolution will not allow the precise location of sulfur at the ultrastructural level as can be done by means of cryo-EFTEM microanalysis. Nevertheless even a 10 μm resolution would permit detection and characterization of the speciation of sulfur and its location at mesoscale within a very short time.
The main advantages of this technique are: (i) its possible application to living samples in their natural environment (in the present study seawater); (ii) its non-destructive nature (interaction of visible light with sample); (iii) its short acquisition time (a few seconds to a few minutes depending on the exciting line power and sulfur concentration); and (iv) its capability to extract chemical species maps or profiles with spatial resolution in the micrometre range.
Materials and Methods
Collection of material
E. dianae (Hopper and Cefalu, 1973) and C. orbicularis (Linné, 1758) were both collected from the tropical seagrass bed of Thalassia testudinum. The tropical site is the ‘îlet cochon’ located in the ‘petit cul de sac marin’ in Guadeloupe (French West Indies, Caribbean).
Bivalves were manually collected, whereas nematodes were extracted by sieving on a 2 mm mesh from a 145 mm diameter core of sediment. The nematodes E. dianae were extracted using a binocular microscope and were re-suspended in filtered seawater (0.22 μm) before analysis.
Electron microscopy (SEM, TEM) techniques
Sample preparation methods
For TEM, individual organisms were fixed for 3 h at 4°C in 4% (w/v) paraformaldehyde in 0.22 μm filtered seawater. After two washes in sterile seawater at room temperature (24°C) and dehydration through an ascending ethanol series, they were embedded in hydrophilic LR White resin (Biovalley) as previously described (Gros and Maurin, 2008). Sections (60 nm) were stained for 30 min in 2% aqueous uranyl acetate and 10 min in 0.1% lead citrate before examination in a Philips 201 microscope running at 75 kV. Semi-thin sections were stained with Toluidine Blue in 1% borax buffer and observed on a Nikon epi80i light microscope.
For SEM, animals were fixed as described above before dehydration in an acetone series and critical-point drying. Samples were then gold coated (sputtering) before observation with a Hitachi SEM S2500 running at 20 kV.
An overall view and histological information was obtained from 7 μm paraffin sections. A gill dissected from one individual, was fixed in Bouin's fluid (Gabe, 1968) for 24 h at room temperature, and was then embedded in paraplast before histological staining according to the method of Frenkiel and Mouëza (1995).
Thin sections from an LR White resin block were investigated with a LEO 912 TEM microscope. An Omega transmission electron microscope (LEO Electron Optics GmbH, Oberkochen, Germany) equipped with a LaB6 source and operated at 120 kV was used. The LEO 912 features an in-column spectrometer [magnetic omega-type electron energy-loss filter (Egerton, 1986; Crozier, 1995)]. PEELS (parallel electron energy loss spectra) were recorded with a cooled slow-scan CCD (charged-coupled device) camera operating in 14-bit mode. The Omega filter was adjusted to obtain an energy resolution of 1.5 eV.
Outline of the technique
Raman spectrometry is based on an inelastic light/matter interaction (Figure 4). It allows us to probe the vibrational energy levels of materials, i.e. molecular or crystalline bonding. The probed material can either be in the gaseous, liquid or solid state. This spectrometry cannot be applied to metallic compounds.
Figure 5 shows the quantum mechanics representation of the main interaction processes of light with a molecule.
The sensitivity of Raman spectrometry is highly dependent on the probed compound and especially on the energy difference, ΔE, between ground and excited electronic levels. When the energy of the exciting photons is far from ΔE, ‘normal’ Raman scattering occurs; intensity represents about 10−6 of the intensity of the incident light, leading to a relatively low sensitivity (Figure 5).
When the energy of the incident photon is close to ΔE, pre-resonance or resonance Raman scattering occurs and the Raman scattering intensity can be enhanced by a factor of 102–103, thus increasing the sensitivity of the method (Long, 1977).
The choice of Raman as the analytical technique used to identify and locate sulfur in low concentrations in biological samples has been encouraged by the fact that sulfur in the S8 structure presents pre-resonance Raman scattering (Herzberg, 1945).
Computer-controlled acquisition and data analysis treatment were developed at GTSI in order to acquire Raman profiles and maps. For this purpose Raman point spectra are acquired along lines in the case of profiles and along several lines covering a whole rectangular area in the case of a mapping (Himmel, 2005).
Quantitative interpretation of the spectra
It has been demonstrated previously (Himmel, 2005) that the probe geometry in the experimental conditions of this work can be represented by a cylinder of 120 μm length by 10 μm diameter. In the case of an homogeneous and transparent sample, the intensity of a characteristic Raman line is linearly related to the analysed volume. In the case of an absorbing sample the intensity of the Raman lines must be corrected by the local attenuation of the exciting line and Raman scattered light. Taking into account the diameter of the nematode, the thickness of its bacterial coat and the exciting line attenuation profile in the sample, Raman spectra treatment will allow us to determine the location of sulfur.
Influence of the light absorption by the various parts of the sample on the Raman scattering intensity
In order to evaluate the absorption of incident laser light as well as the Raman scattered light, we first recorded an optical micrograph of the ectosymbiotic nematode in transmission mode. White light was used for illumination and a green pass band filter was used (514 nm< λ <556 nm) to analyse the transmitted light (Figure 6a).
The intensity of the transmitted light through the bacterial coat, the nematode body (close to the digestive tube) and the body, including the digestive tube, were measured using free Gwyddion software.
Figure 6(b) shows two transmitted light intensity profiles corresponding to lines crossing the nematode surrounded by its bacterial coat and a part of the nematode where the bacterial coat has been removed.
Assuming that the transmitted light intensity is controlled by the Lambert—Beer law:
I0 being the incident light intensity, μ the absorption coefficient, x the thickness of matter crossed by the light and I being the intensity of the transmitted light through the thickness x. We can deduce from the ‘green light’ optical micrograph the absorption coefficient μ corresponding to each part of the sample. The data obtained are summarized in Figure 6(c). Using these coefficients, and applying the Lambert—Beer formula for all contributions of the samples' parts (bacterial coat, nematode body, digestive tube), the exciting line intensity can be calculated point by point.
The frequency of the Raman scattered light of interest (536< λ < 546 nm) lying in the frequency range of the pass band filter used for the recording of the nematode micrograph, the Lambert—Beer equation is also applied to the damping of the Raman scattered light using the absorption coefficients deduced previously (Figure 6c). The full equations are given in the following section.
Mathematical expression of the Raman scattered light intensity for an idealized ectosymbiotic nematode
In order to express the Raman scattered light intensity which takes into account the light absorption by the sample, the Lambert—Beer law equation is applied to an idealized composite sample where the various parts (bacterial coat, nematode body, digestive tube) are considered homogeneous. Figure 7 presents the cross-section of the idealized nematode and the parameters used in the equations.
As previously mentioned, the intensity of light crossing a thickness x of absorbing medium with an absorption coefficient μ is given by the Lambert—Beer law (see above).
Applied to the different parts of the nematode for the incident exciting line and Raman scattered light we can calculate the Raman intensity scattered by the various parts of the sample. The corresponding equations are given below, where: IR is the Raman intensity emitted from the illuminated sample, σ is the Raman cross-section of the studied species (in the present study this is S8), and cis the concentration of S8 in the various parts of the sample.
The thickness crossed by light in the various parts of the sample (Figure 7) are noted: xC (Coat1), xB−xC (Body1), xD−xB (DT, Digestive tube), x′B−xD (Body2) and x′C−x′B (Coat2):
The total Raman intensity detected after a pathway constituted by the succession coat1, body1, digestive tube, body2, coat2 is expressed by the following sum:
For the present study, a Raman microspectrometry is used (HR800 HORIBA Jobin Yvon Microspectrometer), allowing us to obtain high spatial resolution analyses (probe diameter 10 μm). The exciting monochromatic light is the 532 nm wavelength radiation emitted by a 5 W NdYag laser (Spectra-Physics Millennia Pro 5 sJS). The laser power on the sample was limited to 40 mW to avoid sample damage. A 600 lines per mm grating was used, allowing recording of a one shot spectrum over an energy range of 0 cm−1 to 1720 cm−1, the spectrometer entry slit width was 500 μm leading to a resolution of 1.5 cm−1. The confocal hole diameter used was 500 μm leading to a probe length of 400 μm, well adapted to the diameter of the studied samples. The acquisition time was 7 s, to obtain a good signal-to-noise ratio. To prevent damage induced by overheating, samples were maintained in seawater, which constitutes a good cooling isotonic medium for marine invertebrates. The spectra of the seawater are recorded separately and then can be removed from the spectra acquired on the sample in seawater.
Other potential Raman microspectrometry applications in biology
Raman spectrometry permits molecular and structural analysis and can be adapted to function in different atmospheres. In biology this technique can be applied to the identification of mineral phases such as the constituent of mollusc shells or the mineral content of the digestive tube.
At the level of identification of biological molecules and their distribution, Raman spectrometry can mainly be used in the case of molecules presenting high-intensity Raman spectra, i.e. coloured molecules such as metalloproteins or metal chelates.
We would like to thank Y. Bercion and K. Delbé for their technical support.
This work was supported by the Region Guadeloupe the European Social Fund [grant number CAB/AA/JL/No02-9]; and the European Regional Development Fund [grant number A-31-44 présage 6257].