SEARCH

SEARCH BY CITATION

Keywords:

  • acclimation;
  • biodiversity;
  • cell composition;
  • FTIR spectrometry;
  • hierarchical cluster analysis;
  • homeostasis;
  • phytoplankton;
  • species identification

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Conclusions
  6. Acknowledgments
  7. Supporting Information

Fourier transform infrared (FTIR) spectrometry was used to study the spectral features of 12 eukaryotic and two prokaryotic species of microalgae. The algae were cultured in liquid media containing either NO3 or NH4+ as the sole N-source; for the NH4+ treatment, the algae were subjected to short-term (24 h) or long-term (1 month) incubations; for the hypersaline species, cells were also grown in the presence of 2 M NaCl. Over 500 spectra, acquired from at least three distinct cultures for each species, in each growth regime, were subjected to hierarchical cluster analysis (HCA) and were successfully separated according to their taxonomy, showing that the overall spectra were characteristic of each species and that this technique could be fruitfully employed to separate microalgal species living in a similar condition (as would be the case for a natural assemblage). In addition, in most cases, it was possible to differentiate between algae subjected to different growth treatments although belonging to the same species. We also demonstrated that it is possible to accurately identify species and determine the nutritional status of their environment of origin (e.g., N-source), provided that suitable FTIR spectral libraries are available. This study aims to provide the basis for the development of rapid, easy, and inexpensive methods for the evaluation of biodiversity in natural phytoplankton samples and to monitor the water quality of natural environments.

Abbreviations:
FTIR

Fourier transform infrared

HCA

hierarchical cluster analysis

ID test

identity test

The analysis of phytoplankton biodiversity is usually based on microscope counts of cells that are identified on the basis of morphological (e.g., Utermöhl 1958) or molecular (e.g., Caron et al. 2004) features. These procedures are often expensive, tedious, and time consuming and make routine, extensive analyses rather difficult. When counts and identification involve visual analysis of morphological features, moreover, an underestimation of both abundance and biodiversity is not infrequent (e.g., Giovannoni et al. 1990), especially for the least abundant and the smallest organisms (Moon-Van der Staay et al. 2001). Biodiversity is usually associated with water (Harley et al. 2006) and ecosystem (Schlapfer and Schmid 1999) quality. The use of biodiversity as an indicator of sudden and acute changes in water quality requires that the biological assemblages used as signal have a rapid response time. Phytoplankton species composition satisfies this requirement (Niemi et al. 2004). HPLC analysis of pigments is sometimes used to obtain information on changes in phytoplankton composition (Wilhelm et al. 1991, Van Leeuwe et al. 2006). These methods, in addition to having all the technical difficulties usually associated with HPLC, can only provide a very rough idea of species or even genera composition and do not allow reliable estimates of species relative abundance, due to the variability of pigment content per cell (C. Wilhelm, personal communication). The difficulty in obtaining rapid and frequent determinations of phytoplankton species composition has restricted the use of phytoplankton as signal organisms to detect and evaluate environmental change. The possibility of monitoring changes in the species composition of mixed algae assemblages would also represent an exceptional experimental tool for global changes and food web studies (Hays et al. 2005). For all these applications, it is necessary to develop methodologies that allow a very rapid, simple, and inexpensive determination of the taxonomic complexity of phytoplankton assemblages and that provide an equally fast and objective way to identify species.

FTIR spectroscopy is traditionally employed for the analysis of pure compounds or simple mixtures (e.g., Smith 1999, Coates 2000). Its application to whole cells is relatively recent (Naumann et al. 1991, Wong et al. 1991, Kümmerle et al. 1998,Wetzel et al. 1998, Giordano et al. 2001). This approach provides a snapshot of the cellular biochemical composition without the need for complex extraction procedures, and it thereby offers an attractive option for phytoplankton analysis. Since the publication of the first paper on applications of FTIR spectroscopy to whole microalgal cells (Giordano et al. 2001), the use of this methodology has been gaining momentum for physiological and ecophysiological studies, using both traditional (Beardall et al. 2001, Sigee et al. 2002, Stehfest et al. 2005, Montechiaro et al. 2006, Giordano et al. 2007, Jakob et al. 2007, Mecozzi et al. 2007, 2008) and synchrotron IR sources (Heraud et al. 2005, Dean and Sigee 2006, Hirschmugl et al. 2006, Dean et al. 2007).

For a limited number of bacteria (Naumann 2000) and cyanobacteria (Kansiz et al. 1999, Dean and Sigee 2006) species, it has been demonstrated that the FTIR spectra are sufficiently characteristic to allow the identification of species. The species-specificity of FTIR spectra for eukaryotic organisms has only been thoroughly tested in yeasts (Kummerle et al. 1998, Mariey et al. 2001 and references therein). The possibility of distinguishing among eukaryotic algae appears especially challenging, due to the often substantial contribution of chloroplasts to the overall cell composition; because of the complexity and intricacy of the evolutionary trajectories of plastidial endosymbiosis, algae with similar plastids may produce similar spectra in spite of otherwise very distant relationships (Grzebyk et al. 2003, Bhattacharya et al. 2004). On the other hand, similar chloroplasts may have different sizes and be present in different numbers in algae belonging to even closely related species; the plastid contribution to the overall cell spectrum may therefore vary substantially in spite of close taxonomic relationships.

The FTIR signature of algal cells is affected by their nutritional status (Giordano et al. 2001, 2007, Heraud et al. 2005, Montechiaro et al. 2006). This effect offers an additional potential application of FTIR spectroscopy to derive environmental information; in fact, it is theoretically possible to attribute a given FTIR signature to a certain nutritional status and thus use whole-cell spectra to monitor the trophic condition of a given environment. On these bases, the aim of this work was to verify if: (1) FTIR spectra of microalgae are species specific; (2) environmental conditions can, at least to some extent, be determined from FTIR spectra; (3) it is possible to identify unknown species using an appropriate FTIR library.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Conclusions
  6. Acknowledgments
  7. Supporting Information

Cultures.  Fourteen algal species, belonging to five different divisions and comprising common phytoplankters of hypersaline, marine, and freshwater environments were used for this study (see Table 1, which contains taxonomic authors and strain numbers). All species were grown in axenic, semicontinuous cultures, at a photon flux density of 100 μmol · m−2 · s−1 (PAR), at 20°C. To mimic the situation in which a variety of species is present in a natural assemblage, all saltwater species (included the halotolerant Dunaliella spp.) were cultured in the same ESAW medium, pH 8.2 (Berges et al. 2001), and all freshwater species were grown in BG11 medium, pH 7.1 (Rippka et al. 1979). Both seawater and freshwater species were also transferred from standard ESAW and BG11 (containing NaNO3 as the N source) to media in which NaNO3 had been substituted with an equal molar concentration of NH4Cl. Analyses were conducted after the cells were incubated for 24 h and 1 month in NH4+-medium. For the hypersaline species, Dunaliella salina and Dunaliella parva, spectra were also obtained from cells cultured for at least 1 month in a medium that differed from ESAW because its salinity was increased with the addition of NaCl to a final concentration of 2 M.

Table 1.   Taxonomy of the 14 species used in this study.
PhylumClassOrderFamilySpeciesStrain
  1. N.A., not available.

  2. aP. reticulatum was isolated and kindly donated by Dr. Franca Guerrini, University of Bologna, Italy.

  3. bS. microadriaticum was isolated from the sea fan Eunicella singularis collected in Portofino, Italy, and was kindly donated by Dr. Carlo Cerrano, University of Genova, Italy.

CyanobacteriaCyanophyceaeOscillatorialesPhormidiaceaePhormidium autumnale (C. Agardh) Trevisan ex GomontCCAP 1462/10
SynechococcalesMerismopediaceaeSynechocystis sp.PCC 6803
ChlorophytaChlorophyceaeVolvocalesChlamydomonaceaeChlamydomonas reinhardtii P. A. Dang.CC 125
DunaliellaceaeDunaliella parva W. LercheCCAP 19/9
Dunaliella salina (Dunal) Teodor.CCAP 19/25
TrebouxiophyceaeChlorellalesChlorellaceaeChlorella marina ButcherCCAP 211/27
PrasinophyceaeChlorodendralesChlorodendraceaeTetraselmis suecica (Kylin) ButcherPCC 305
BacillariophytaCoscinodiscophyceaeThalassiosiralesThalassiosiraceaeThalassiosira weissflogii (Grunow) G. A. Fryxell et HaslePCC 9115
SkeletonemaceaeSkeletonema marinoi Sarno et ZingoneCCMP 2092
BacillariophyceaeNaviculalesPhaeodactylaceaePhaeodactylum tricornutum BohlinCCAP 1052/1A
MyzozoaDinophyceaePeridinialesGonyaulacaceaeProtoceratium reticulatum (Clap. et J. Lachm.) BütschliPRA 0206a
SuessialesSymbiodiniaceaeSymbiodinium microadriaticum Freud.N.A.b
HaptophytaPrymnesiophyceaeIsochrysidalesNöelhaerabdaceaeEmiliania huxleyi (Lohmann) W. H. Hay et H. MohlerCCAP 920/11
IsochrysidaceaeIsochrysis galbana ParkeCCMP 1323

Cell densities were determined via direct counts using a Burker hemocytometer (Larsson et al. 1978) or, for the larger Protoceratium reticulatum, a Sedwick-Rafter chamber (McAlice 1971). Since it was not possible to reliably determine cell density for the filamentous Phormidium autumnale (Montechiaro et al. 2006), the abundance of this organism was estimated from its chl content measured according to Montechiaro and Giordano (2006). Motile species were fixed prior to counting by the addition of 3 μL of Lugol’s solution (Throndsen 1978) to 1 mL of algal suspension.

Infrared analysis.  Cell deposition:  Preliminary trials were performed to determine the density of cell suspensions necessary to produce spectra with an optimal signal-to-noise ratio without band saturation. Cells were harvested by centrifugation at 1,800g for 15 min with a ALC 4235A Centrifuge (ALC International, Milan, Italy), except in the case of P. reticulatum, whose cells tended to break when centrifuged and were thus collected by filtration through nylon filters with a pore size of 11 μm (Millipore, Billerica, MA, USA). Cells were then washed twice with an iso-osmotic solution of ammonium formate to minimize medium carryover and thus optical scattering effect on the spectra. After the final wash, cells were resuspended in a volume of ammonium formate solution such that the appropriate number of cells was contained in 50 μL. This volume was deposited on silicon windows (Crystran Ltd., Poole, UK) and desiccated in an oven at 60°C for 3 h; the samples were kept in a desiccator until analyzed. The experimental blanks were prepared depositing 50 μL of ammonium formate solution on the windows and subjecting these windows to the same treatment described for the samples. No trace of ammonium formate was detected in the blank spectra.

Spectral acquisition:  Spectra were collected with a Tensor 27 FTIR spectrometer (Bruker Optics, Ettlingen, Germany). The Bruker system was controlled by an IBM-compatible PC running the software OPUS version 6.0 (Bruker Optics 2006). The absorbance spectra were collected between 3,500 cm−1 and 500 cm−1, at a spectral resolution of 4 cm−1, with 32 scans co-added and averaged. A Blackman-Harris three-term apodization function was used, with a zero-filling factor of 2. To minimize differences between spectra due to baseline shifts, and spectra deformation due to scattering, the spectra were baseline corrected using the “Scattering correction” algorithm within the OPUS 6.0 software, using 200 baseline points and excluding the CO2 bands. All spectra were normalized to the amide I band at ∼1,655 cm−1 to account for any difference in the deposit thickness (Giordano et al. 2001).

Spectral analyses:  More than 30 spectra for each species, for a total of 500 spectra, were analyzed. Spectra were analyzed using the statistical package of OPUS 6.0. Peak attribution was performed according to Giordano et al. (2001). The contribution of the various absorption bands to cell composition was estimated relative to the amide I band.

Spectra similarities were compared by HCA using the function “Cluster Analysis” of the OPUS 6.0 software. Prior to the analysis, spectra were converted to first derivative using the Savitzky-Golay algorithm (Kansiz et al. 1999) and subjected to vector normalization (e.g., Lasch et al. 2004). HCA was performed on the following spectral ranges: 3,500–500 cm−1, 3,500–900 cm−1, and 1,800–900 cm−1, the latter being the portion of the spectrum usually richer of biochemical information. Dendrograms were obtained with both the Ward’s method (Ward 1963) and the single linkage method (Everitt 1993), on the basis of Euclidean distances.

Identification of spectra was carried out by binary isolate-to-reference spectra similarity comparisons (identity test), which provides a hit-list indicating the level of similarity of the reference spectra to the spectrum to be identified (Oberreuter et al. 2002). This procedure was carried out using the “Identity Test” function of the OPUS 6.0 software. The reference library was constituted by 168 spectra, representing all species cultured in NO3 or incubated in NH4+ for 24 h. These spectra were organized in 14 classes corresponding to the 14 species in Table 1. Each class included two subclasses, one composed of spectra from NO3-grown cells and the other with spectra obtained from cells incubated for 24 h in medium containing NH4+ as the sole N source. For the two hypersaline Dunaliella species, one additional subclass was created for the cells grown at 2 M NaCl. For all subclasses, the Euclidean distance between the mean spectrum and each spectrum in the subclass was calculated. The largest distance (worst case) determined the value of the threshold of that subclass. The threshold was extended by a value that was conservatively set at 0.25 of one standard deviation, to allow for the possibility of a slightly larger subclass heterogeneity than that represented in the library. The threshold values for all subclasses are shown in Table S1 in the supplementary material.

The ID test was carried out on 133 spectra taken from those not used to construct the library. The distances between a spectrum to be identified and the subclasses’ mean spectra were the hit qualities for that spectrum. A hit quality smaller than the threshold value of a given subclass (i.e., the spectrum was closer to the mean spectrum than the threshold distance) indicated that the spectrum was a member of that subclass. A univocal subclass match was required for a positive identification. When a spectrum fell within the boundaries of two subclasses belonging to the same class (non-univocal identification, but consistent attribution to one species), it was identified at the class level.

Results and discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Conclusions
  6. Acknowledgments
  7. Supporting Information

Features of FTIR spectra obtained from the different species under various culture conditions.  Representative spectra for all species cultured in the various growth conditions are shown in Figure 1. The description of the results that follows refers to the area of the relevant peaks with respect to the amide I peak used for normalization. Qualitative differences among spectra are often obvious. For instance, the typical silicate peak characterizes the spectra of diatoms (Giordano et al. 2001); however, the notoriously less silicified diatom Phaeodactylum tricornutum (Francius et al. 2008 and references therein) showed a much smaller silicate peak, as compared to the other diatoms analyzed. Two haptophytes were analyzed; both of them showed a low absorbance in the 1,200–900 cm−1 bands, due to carbohydrate functional groups (Giordano et al. 2001), with a somewhat larger relative intensity in this spectral region for Isochrysis galbana. In the spectra obtained from the marine dinoflagellate P. reticulatum, the peak intensities in the 1,200–900 cm−1 range were very high, as compared to the amide I peak; however, this was not a trait shared by the other dinoflagellate, Symbiodinium microadriaticum (notice that, for this study, free-living cells of the coral symbiont S. microadriaticum were used). The low carbohydrate to protein ratio in free-living S. microadriaticum is in itself an interesting finding, and it calls for further studies to determine the extent that resource allocation in free-living cells differs from that in symbiotic cells (Smith et al. 2005, Loram et al. 2007). Green algae showed high relative absorbance in the spectral window 1,200–900 cm−1; only Chlorella marina, among the green algae, had lower absorbance in this region than in the amide I band. It is noteworthy that the Chlamydomonas reinhardtii strain used (Table 1) did not contain a functional gene encoding for nitrate reductase (nia1) and was therefore barely surviving in NO3-containing medium, possibly relying on the minimal amount of NH4+ present in the BG11 medium as ammonium ferric citrate and/or on N derived from amino acid released by dead cells (Ermilova et al. 2007). The difficulty of this algae in acquiring N is confirmed by the high ratio of the 1,740 cm−1 feature, usually attributed to lipids, and of the carbohydrates peaks relative to the amide I peak (Giordano et al. 2001); it is not uncommon that, in the presence of a C:N internal ratio higher than optimal, the lipid and carbohydrate pools become greater with respect to the protein pool (amides absorption bands; Huppe and Turpin 1994, Giordano et al. 2001). The two cyanobacteria showed very different spectra; the relative contribution of peaks in the 1,200–900 cm−1 region was substantially larger in Synechocystis PCC6803 than in P. autumnale.

image

Figure 1.  Representative FTIR spectra of Phaeodactylum tricornutum, Skeletonema marinoi, Thalassiosira weissflogii, Emiliania huxleyi, Isochrysis galbana, Protoceratium reticulatum, Symbiodinium microadriaticum, Chlamydomonas reinhardtii, Chlorella marina, Dunaliella parva, Dunaliella salina, Tetraselmis suecica, Phormidium autumnale, Synechocystis sp. For each species, the spectra obtained from cells cultured in NO3, incubated in the presence of NH4+ for 24 h, or incubated in medium at higher salinity (2 M NaCl, only for Dunaliella spp.) are shown. FTIR, Fourier transform infrared.

Download figure to PowerPoint

The spectral differences between cells cultured under different N and salinity growth regimes were various. For instance, D. parva, C. marina, and P. autumnale showed no change in their FTIR spectra after transfer to NH4+; the stability of the macromolecular pools of P. autumnale and D. parva to environmental perturbations has been previously reported in the literature (Montechiaro et al. 2006, Giordano et al. 2007). All other species showed appreciable alterations in their spectra, although to different extents, when culture conditions were modified; this finding is indicative of a higher compositional plasticity and possibly of a greater tendency to respond by acclimation to changes in external conditions (Raven and Geider 2003, Fig. 1). It is noteworthy that even closely related organisms, such as the congeneric species D. salina and D. parva, which have similar ecological, functional, and morphological characteristics, adopt different strategies in response to environmental change, at least with respect to their macromolecular composition. Even the salinity change, to which both species adjust in a similar way (i.e., by the synthesis of glycerol to osmotic equilibrium with the medium; Ben-Amotz et al. 2009), caused profound changes in D. salina spectra and essentially no modifications in D. parva spectra (Fig. 1).

HCA and species specificity of spectra.  To discriminate among different species without the need of a spectrum by spectrum comparison, a simple HCA was applied. The spectra from all 14 species clustered independently from each other, regardless of the N-source in the growing medium (separation at the class level, according to the terminology used for the ID test; see Materials and Methods). This separation was obtained for the spectral ranges 3,500–500 cm−1, 3,500–900 cm−1, and 1,800–900 cm−1, applying both the Ward’s method and the single linkage method (the dendrogram in Fig. 2 was obtained for the range 1,800–900 cm−1 with the Ward’s method; Lasch et al. 2004; see Figs. S1–S6 in the supplementary material for dendrograms comprising all N nutritional growth regimes, generated with both methods, for all ranges).

image

Figure 2.  Dendrogram showing the degree of similarity for all 14 species tested (Phormidium autumnale, Synechocystis PCC 6803, Dunaliella parva, Dunaliella salina, Chlamydomonas reinhardtii, Chlorella marina, Tetraselmis suecica, Thalassiosira weissflogii, Skeletonema marinoi, Phaeodactylum tricornutum, Protoceratium reticulatum, Symbiodinium microadriaticum, Emiliania huxleyi, Isochrysis galbana), cultured in NO3, in medium at higher salinity (2 M NaCl), or incubated in the presence of NH4+ for 24 h. The dendrogram was generated from spectra collected in the wavenumber range 1,800–900 cm−1 and was constructed according to the method by Ward (1963).

Download figure to PowerPoint

The HCA analysis also separated the spectra of cells of the same species exposed to different conditions (separation at the subclass level). The only exceptions were P. autumnale and D. parva, which are known to be little responsive to changes in growth regimes (Montechiaro et al. 2006, Giordano et al. 2007). In most cases, the spectra from the same species subjected to different culture conditions, although separate, clustered together at the subsequent nesting level, except for D. salina and C. reinhardtii, which are reported to be especially plastic species (Giordano and Bowes 1997, Grossman et al. 2007). The various species were also convincingly separated, if the cells were cultured in the presence of NH4+ for 1 month (Figs. S1–S6). The effect of this treatment at the level of subclasses is more variable: some algae accentuated the features that characterized their spectra after 24 h in NH4+, so that the separation between subclasses (NO3 vs. NH4+-grown cells) became more obvious (e.g., Tetraselmis suecica, C. marina, Skeletonema marinoi, S. microadriaticum, D. parva); other species appeared to become more similar to the NO3-grown cells (e.g., Synechocystis sp., Emiliania huxleyi, Thalassiosira weissflogii, P. reticulatum, D. salina), making it more difficult to differentiate between subclasses (Figs. S1–S6). These results presumably reflect the resilience (homeostasis) of cell composition, or the tendency of cells to acclimate to culture conditions (Raven and Geider 2003), and can provide fundamental information on the response strategies of different algal species.

With rare exceptions, species clustering is approximately compatible with known taxonomic relationships. An interesting exception is given by P. tricornutum, which did not cluster with the other two diatoms and was instead close to the haptophytes (Fig. 2). This association is presumably the consequence of the typical low silicate content of P. tricornutum cell walls (Francius et al. 2008) and of the relatively low carbohydrate content (evidenced by our spectroscopic measurements) of this species that make its spectra resemble those of the haptophytes (Fig. 1). This example suggests that, although HCA of FTIR spectra from whole microalgal cells has clear taxonomic significance (all species are separated, and in most cases, related organisms are clustered close to each other), it cannot be used to establish phylogenetic relationship due to the impact that specific compositional features (e.g., amount of silicate in the frustule) can have on the outcome of the analysis. The clustering pattern of the species used for this study does not seem to reflect chloroplast phylogeny either; although most green algae cluster together, C. marina and NH4+-grown D. salina and C. reinhardtii are grouped together with the chl a/c dinoflagellates, and the cyanobacteria cluster with the haptophytes and P. tricornutum.

Species identification.  Since the HCA showed that, at least with the species used for this study, FTIR spectra from both eukaryotic and prokaryotic algae were highly species-specific, it should be possible to use the spectra for taxonomic identification of species, provided that an appropriate spectral library is available. Using the library constructed as described in the Materials and Methods, we attempted to verify if this was indeed possible. The ID test was carried out on all 14 species, on both NO3-grown cells and cells incubated in NH4+ for 24 h and, for the Dunaliella species, also on cells cultured at high salinity (Fig. 3). The ID test afforded the correct identification at the class level (species) for 16.5% of the spectra analyzed (22 of 133; 20 of these spectra had the correct subclass as their best hit) and for the 66.2% (88 spectra of 133) at the subclass level (species and growth condition); only 23 spectra of 133 (17.3%) were not identified by the OPUS 6.0 Identity Test routine (Fig. 3a). Among the spectra that did not pass the ID test, 65.2% were not identified because their hit values were outside of all thresholds, whereas in 34.8% of the cases, the hit values fell within more than one threshold (non-univocal identification). However, a closer examination of the best hits shows that 30.4% of the spectra (7 of 23) that did not satisfy the stringent requirements of the ID test had their best hit with the correct class (species), and 69.6% (16 spectra of 23) had their higher similarity with the correct subclass (species and condition; Fig. 3b). With this more permissive interpretation of our results (i.e., on the basis of the best hits, also considering the best hits of the spectra identified at the class level), 100% of the total number of spectra analyzed would be attributed to the correct species, and 93.2% (124 spectra) would be assigned to the correct species and culture condition.

image

Figure 3.  Percentages of success and failure of spectra identification by the identity test. Panel A shows the overall percentages of successful identification at the level of species (class) and of species and culture condition (subclass); the striped area indicates the percentage of spectra that did not pass the identity test. Panel B shows a more detailed analysis of the spectra not identified by the identity test, indicating whether their best hit was correct at the class or subclass level. No spectrum had a best hit with the wrong species.

Download figure to PowerPoint

In general, homeostatic species were identified more confidently (lower hit value) at the level of classes, due to their modest compositional variability (Table S1). However, their growth conditions were correctly detected with a larger degree of uncertainty than in the case of more plastic species, since cell composition tended to be less responsive to changes in the external environment (Figs. S1–S6).

We intentionally created libraries with different qualitative and quantitative compositions and observed that the lower limit for the number of spectra that constitute a library subclass depended on the plasticity of the organism that constitutes that subclass; very plastic organisms require a larger number of spectra for a confident identification than organisms with lower compositional heterogeneity in their populations (data not shown). On the other hand, the inclusion of outliers can increase the boundary of subclasses to the point of making separations impossible. It is thus crucial that the physiology of the algae from which the library spectra are derived is carefully characterized, so that, especially at the level of subclass, identification is precise. Although the rate of success of the ID test was very high, in this study (Fig. 3), we acknowledge the fact that, in much larger libraries, a greater degree of non-univocal attribution of spectra to subclasses may occur and/or the identification of unknown specimens may require a more direct intervention of the operator to sort out the outcome of the identity test.

Conclusions

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Conclusions
  6. Acknowledgments
  7. Supporting Information

In the light of the above results, FTIR spectroscopy appears to have considerable potential for biodiversity studies. Spectra obtained from organisms that were both very distant and very close in terms of taxonomy, physiology, morphology, and ecology were used in this study, and all of them were clearly separated by applying an intentionally simple and straightforward HCA procedure. Using an IR microscope connected to an FTIR spectrometer, the analysis of spectra can be performed rapidly on natural samples, substantially reducing the time required for a determination of species composition. Furthermore, the development of more affordable and faster focal plane array detectors with high resolution (Levenson et al. 2006, Heraud et al. 2007) or of microscope stages controlled by software for shape recognition may allow the concomitant and/or automatic acquisition of multiple spectra, with a dramatic simplification of phytoplankton biodiversity evaluation.

In most cases, it was also possible to discern the growth conditions of the cells from which the spectra were collected. The degree of success in the determination of the culture conditions is largely dependent on the compositional plasticity of the species. In practical terms, the selection of appropriate signal species (i.e., very responsive species among those naturally present in a given area) may allow the use of FTIR spectroscopy as an easy, fast, and inexpensive methodology to monitor natural waters. The appropriate choice of signal species will also determine the time frame within which this approach can be applied: some species respond to external change by rapidly and greatly changing their composition; others do so only after more prolonged exposure to the exogenous stimuli (e.g., see data for algae incubated for 24 h and 30 d in the presence of NH4+, in this study). There are also species that undergo substantial changes when confronted with a variation in nutrient availability, then revert to their initial composition (Giordano et al. 2007). In some cases, change may cause genotype selection; even these adaptive responses would be detected by FTIR spectroscopy, as long as they lead to a modification in the relative abundance of IR-absorbing pools. Because species responses depend on the type and intensity of the stimuli (Montechiaro et al. 2006), the parameters to be monitored would also affect the choice of the signal organism.

Identification of unknown species also appears possible via FTIR spectroscopy, although library structure becomes a critical factor in the success rate of identification.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Conclusions
  6. Acknowledgments
  7. Supporting Information

We wish to thank Prof. John Raven and Prof. Christian Wilhelm for their revisions of the manuscript. Our gratitude also goes to Dr. Frank Vogt for his advice on the chemometric methods.

  • Beardall, J., Berman, T., Heraud, P., Kadiri, M. O., Light, B. R., Patterson, G., Roberts, S., et al. 2001. A comparison of methods for detection of phosphate limitation in microalgae. Aquat. Sci. 63:10721.
  • Ben-Amotz, A., Polle, J. E. W. & Subba Rao, D. V. 2009. The Alga Dunaliella: Biodiversity, Physiology, Genomics & Biotechnology. Science Publishers, Enfield, New Hampshire (in press).
  • Berges, J. A., Franklin, D. J. & Harrison, P. J. 2001. Evolution of an artificial seawater medium: improvements in enriched seawater, artificial water over the last two decades. J. Phycol. 37:113845.
  • Bhattacharya, D., Yoon, H. S. & Hackett, J. D. 2004. Photosynthetic eukaryotes unite: endosymbiosis connects the dots. BioEssays 26:5060.
  • Bruker Optics. 2006. OPUS 6.0 Reference Manual. Bruker Optics, Ettlingen, Germany, 456 pp.
  • Caron, D. A., Countway, P. D. & Brown, M. V. 2004. The growing contributions of molecular biology and immunology to protistan ecology: molecular signatures as ecological tools. J. Eukaryot. Microbiol. 51:3848.
  • Coates, J. 2000. Interpretation of infrared spectra, a practical approach. In Meyers, R. A. [Ed.] Encyclopedia of Analytical Chemistry. John Wiley & Sons, Chichester, UK, pp. 1081537.
  • Dean, A. P., Martin, M. C. & Sigee, D. C. 2007. Resolution of codominant phytoplankton species in a eutrophic lake using synchrotron-based Fourier transform infrared spectroscopy. Phycologia 46:1519.
  • Dean, A. P. & Sigee, D. C. 2006. Molecular heterogeneity in Aphanizomenon flos-aquae and Anabaena flos-aquae (Cyanophyta): a synchrotron-based Fourier-transform infrared study of lake micro-populations. Eur. J. Phycol. 41:20112.
  • Ermilova, E. V., Nikitin, M. M. & Fernandez, E. 2007. Chemotaxis to ammonium/methylammonium in Chlamydomonas reinhardtii: the role of transport systems for ammonium/methylammonium. Planta 226:132332.
  • Everitt, B. S. 1993. Cluster Analysis, 3rd ed. Halsted Press, New York, 170 pp.
  • Francius, G., Tesson, B., Dague, E., Martin-Jézéquel, V. & Dufrêne, Y. F. 2008. Nanostructure and nanomechanics of live Phaeodactylum tricornutum morphotypes. Environ. Microbiol. 10:134456.
  • Giordano, M. & Bowes, G. 1997. Gas exchange and allocation in Dunaliella salina cells in response to the N source and CO2 concentration used for growth. Plant Physiol. 115:104956.
  • Giordano, M., Kansiz, M., Heraud, P., Beardall, J., Wood, B. & McNaughton, D. 2001. Fourier transform infrared spectroscopy as a novel tool to investigate changes in intracellular macromolecular pools in the marine microalga Chaetoceros muellerii (Bacillariophyceae). J. Phycol. 37:2719.
  • Giordano, M., Norici, A., Gilmour, D. J. & Raven, J. A. 2007. Physiological responses of the green alga Dunaliella parva (Volvocales, Chlorophyta) to controlled incremental changes in the N source. Funct. Plant Biol. 34:110.
  • Giovannoni, S. J., Britschgi, T. B., Moyer, C. L. & Field, K. G. 1990. Genetic diversity in Sargasso Sea bacterioplankton. Nature 345:603.
  • Grossman, A. R., Croft, M., Gladyshev, V. N., Merchant, S. S., Posewitz, M. C., Prochnik, S. & Spalding, M. H. 2007. Novel metabolism in Chlamydomonas through the lens of genomics. Curr. Opin. Plant Biol. 10:1908.
  • Grzebyk, D., Schofield, O., Vetriani, C. & Falkowski, P. G. 2003. The mesozoic radiation of eukaryotic algae: the portable plastid hypothesis. J. Phycol. 39:25967.
  • Harley, C. D. G., Hughes, A. R., Hultgren, K. M., Miner, B. G., Sorte, C. J. B., Thornber, C. S., Rodriguez, L. F., Tomanek, L. & Williams, S. L. 2006. The impacts of climate change in coastal marine systems. Ecol. Lett. 9:22841.
  • Hays, G. C., Richardson, A. J. & Robinson, C. 2005. Climate change and marine plankton. Trends Ecol. Evol. 20:33744.
  • Heraud, P., Caine, S., Sanson, G., Gleadow, R., Wood, B. R. & McNaughton, D. 2007. Focal plane array infrared imaging: a new way to analyse leaf tissue. New Phytol. 173:21625.
  • Heraud, P., Wood, B. R., Tobin, M. J., Beardall, J. & McNaughton, D. 2005. Mapping of nutrient-induced biochemical changes in living algal cells using synchrotron infrared microspectroscopy. FEMS Microbiol. Lett. 249:21925.
  • Hirschmugl, C., Bayarri, Z.-E., Bunta, M., Holt, J. B. & Giordano, M. 2006. Analysis of nutritional status of algae by Fourier transform infrared chemical imaging. Infrared Phys. Techn. 49:5763.
  • Huppe, H. C. & Turpin, D. H. 1994. Integration of carbon and nitrogen metabolism in plant and algal cells. Annu. Rev. Plant Physiol. Plant Mol. Biol. 45:577607.
  • Jakob, T., Wagner, H., Stehfest, K. & Wilhelm, C. 2007. A complete energy balance from photons to new biomass reveals a light- and nutrient-dependent variability in the metabolic costs of carbon assimilation. J. Exp. Bot. 58:210112.
  • Kansiz, M., Heraud, P., Wood, B., Burden, F., Beardall, J. & McNaughton, D. 1999. Fourier transform infrared microspectroscopy and chemometrics as a tool for the discrimination of cyanobacterial strains. Phytochemistry (Oxf.) 52:40717.
  • Kümmerle, K., Scherer, S. & Seiler, H. 1998. Rapid and reliable identification of food-borne yeasts by Fourier-transform infrared spectroscopy. Appl. Environ. Microbiol. 64:220714.
  • Larsson, K., Weibull, C. & Cronberg, G. 1978. Comparison of light and electron microscopic determinations of the number of bacteria and algae in lake water. Appl. Environ. Microbiol. 35:397404.
  • Lasch, P., Haensch, W., Naumann, D. & Diem, M. 2004. Imaging of colorectal adenocarcinoma using FT-IR microspectroscopy and cluster analysis. Biochim. Biophys. Acta 1688:7686.
  • Levenson, E., Lerch, P. & Martin, M. C. 2006. Infrared imaging: synchrotrons vs. arrays, resolution vs. speed. Infrared Phys. Techn. 49:4552.
  • Loram, J. E., Trapido-Rosenthal, G. & Douglas, A. E. 2007. Functional significance of genetically different symbiotic algae Symbiodinium in a coral reef symbiosis. Mol. Ecol. 16:484957.
  • Mariey, L., Signolle, J. P., Amiel, C. & Travert, J. 2001. Discrimination, classification, identification of microorganisms using FTIR spectroscopy and chemometrics. Vibrat. Spectrosc. 26: 1519.
  • McAlice, B. J. 1971. Phytoplankton sampling with the Sedgwick-Rafter cell. Limnol. Oceanogr. 16:1928.
  • Mecozzi, M., Onorati, F., Oteri, F. & Sarni, A. 2008. Characterization of a bioassay using the marine alga Dunaliella tertiolecta associated with spectroscopic (visible and infrared) detection. Int. J. Environ. Poll. 32:10420.
  • Mecozzi, M., Pietroletti, M. & Di Mento, R. 2007. Application of FTIR spectroscopy in ecotoxicological studies supported by multivariate analysis and 2D correlation spectroscopy. Vibrat. Spectrosc. 44:22835.
  • Montechiaro, F. & Giordano, M. 2006. Effect of prolonged dark incubation on pigments and photosynthesis of the cave-dwelling cyanobacterium Phormidium autumnale (Oscillatoriales, Cyanobacteria). Phycologia 45:70410.
  • Montechiaro, F., Hirschmugl, C. J., Raven, J. A. & Giordano, M. 2006. Homeostasis of cell composition during prolonged darkness. Plant Cell Environ. 29:2198204.
  • Moon-Van der Staay, S. Y., DeWachter, R. & Vaulot, D. 2001. Oceanic 18S rDNA sequences from picoplankton reveal unsuspected eukaryotic diversity. Nature 409:60710.
  • Naumann, D. 2000. Infrared spectroscopy in microbiology. In Meyers, R. A. [Ed.] Encyclopedia of Analytical Chemistry. John Wiley & Sons, Chichester, UK, pp. 10231.
  • Naumann, D., Helm, D. & Labischinski, H. 1991. Microbiological characterizations by FTIR spectroscopy. Nature 351:812.
  • Niemi, G., Wardrop, D., Brooks, R., Anderson, S., Brady, V., Paerl, H., Rakocinski, C., Brouwer, M., Levinson, B. & McDonald, M. 2004. Rationale for a new generation of indicators for coastal waters. Environ. Health Persp. 112:97986.
  • Oberreuter, H., Seiler, H. & Scherer, S. 2002. Identification of coryneform bacteria and related taxa by Fourier-transform infrared (FT-IR) spectroscopy. Int. J. Syst. Evol. Microbiol. 52:91100.
  • Raven, J. A. & Geider, R. J. 2003. Adaptation, acclimation and regulation in algal photosynthesis. In Larkum, A. W. D., Douglas, S. E. & Raven, J. A. [Eds.] Photosynthesis in Algae. Kluwer Academic Publishers, Dordrecht, the Netherlands, pp. 385412.
  • Rippka, R., Deruellas, J., Waterbury, J. B., Herdman, M. & Stainer, R. Y. 1979. Generic assignments, strain histories and properties of pure cultures of cyanobacteria. J. Gen. Microbiol. 111:161.
  • Schlapfer, F. & Schmid, B. 1999. Ecosystem effects of biodiversity: a classification of hypotheses and exploration of empirical results. Ecol. Appl. 9:893912.
  • Sigee, D. C., Dean, A., Levado, E. & Tobin, M. J. 2002. Fourier-transform infrared spectroscopy of Pediastrum duplex: characterization of a micro-population isolated from a eutrophic lake. Eur. J. Phycol. 7:1926.
  • Smith, B. 1999. Infrared Spectral Interpretation, A Systematic Approach. CRC Press, Boca Raton, Florida, 288 pp.
  • Smith, D. J., Suggett, D. J. & Baker, N. 2005. Is photoinhibition of zooxanthellae photosynthesis the primary cause of thermal bleaching in corals? Glob. Change Biol. 11:111.
  • Stehfest, K., Toepel, J. & Wilhelm, C. 2005. The application of micro-FTIR spectroscopy to analyze nutrient stress-related changes in biomass composition of phytoplankton algae. Plant Physiol. Biochem. (Paris) 43:71726.
  • Throndsen, J. 1978. Preservation and storage. In Sournia, A. [Ed.] Phytoplankton Manual. UNESCO, Paris, pp. 6974.
  • Utermöhl, H. 1958. Toward the improvement of the quantitative phytoplankton method. Mitt. Int. Ver. Limnol. 9:138.
  • Van Leeuwe, M. A., Villerius, L. A., Roggeveld, J., Visser, R. J. W. & Stefels, J. 2006. An optimized method for automated analysis of algal pigments by HPLC. Mar. Chem. 102:26775.
  • Ward, J. H., Jr. 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 58:23644.
  • Wetzel, D. L., Eilert, A. J., Pietrzak, L. N., Miller, S. S. & Sweat, J. A. 1998. Ultraspatially-resolved synchrotron infrared microspectroscopy of plant tissue in situ. Cell. Mol. Biol. 44:14567.
  • Wilhelm, C., Rudolph, I. & Renner, W. 1991. A quantitative method based on HPLC-aided pigment analysis to monitor structure and dynamics of the phytoplankton assemblage – a study from Lake Meerfelder (Eifel, Germany). Arch. Hydrobiol. 123:2135.
  • Wong, P. T. T., Wong, R. K., Caputo, T. A., Godwin, T. A. & Rigas, B. 1991. Infrared spectroscopy of exfoliated human cervical cells: evidence of extensive structural changes during carcinogenesis. Proc. Natl. Acad. Sci. U.S.A. 88:1098892.

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results and discussion
  5. Conclusions
  6. Acknowledgments
  7. Supporting Information

Figure S1. Dendrogram showing the degree of similarity for all 14 species tested, grown in the presence of either NO3 or NH4+ for several generations (not <30 d) or incubated in the presence of NH4+ for 24 h, after transfer from NO3. The dendrogram was generated from spectra collected in the wavenumber range 1,800–900 cm−1 and was constructed according to the method by Ward (1963). The color code of lines is the following: ——— green algae; ——— diatoms; ——— haptophytes; ——— dinoflagellates; ——— cyanobacteria.

Figure S2. Dendrogram showing the degree of similarity for all 14 species tested, grown in the presence of either NO3 or NH4+ for several generations (not <30 d) or incubated in the presence of NH4+ for 24 h, after transfer from NO3. The dendrogram was generated from spectra collected in the wavenumber range 1,800–900 cm−1 and was constructed according to the single linkage method (Everitt 1993). The color code of lines is the following: —— green algae; ——— diatoms; ——— haptophytes; ——— dinoflagellates; ——— cyanobacteria.

Figure S3. Dendrogram showing the degree of similarity for all 14 species tested, grown in the presence of either NO3 or NH4+ for several generations (not <30 d) or incubated in the presence of NH4+ for 24 h, after transfer from NO3. The dendrogram was generated from spectra collected in the wavenumber range 3,500–900 cm−1 and was constructed according to the method by Ward (1963). The color code of lines is the following: ——— green algae; ——— diatoms; ——— haptophytes; ——— dinoflagellates; ——— cyanobacteria.

Figure S4. Dendrogram showing the degree of similarity for all 14 species tested, grown in the presence of either NO3 or NH4+ for several generations (not <30 d) or incubated in the presence of NH4+ for 24 h, after transfer from NO3. The dendrogram was generated from spectra collected in the wavenumber range 3,500–900 cm−1 and was constructed according to the single linkage method (Everitt 1993). The color code of lines is the following: ——— green algae; ——— diatoms; ——— haptophytes; ——— dinoflagellates; ——— cyanobacteria.

Figure S5. Dendrogram showing the degree of similarity for all 14 species tested, grown in the presence of either NO3 or NH4+ for several generations (not <30 d) or incubated in the presence of NH4+ for 24 h, after transfer from NO3”. The dendrogram was generated from spectra collected in the wavenumber range 3,500–500 cm−1 and was constructed according to the method by Ward (1963). The color code of lines is the following: ——— green algae; ——— diatoms; ——— haptophytes; ——— dinoflagellates; ——— cyanobacteria.

Figure S6. Dendrogram showing the degree of similarity for all 14 species tested, grown in the presence of either NO3 or NH4+ for several generations (not <30 d) or incubated in the presence of NH4+ for 24 h, after transfer from NO3. The dendrogram was generated from spectra collected in the wavenumber range 3,500–500 cm−1 and was constructed according to the single linkage method (Everitt 1993). The color code of lines is the following: ——— green algae; ——— diatoms; ——— haptophytes; ——— dinoflagellates; ——— cyanobacteria.

Table S1. ID test results. For each spectrum analyzed, the level of identification (class and/or subclass), the hit quality with the indication of the best hit species, the threshold and confidence level are shown. In the case of unsuccessful identification, the best hit and the reason for the negative result of the ID are reported. The green color indicates identification to the subclass level; the blue color indicates identification to the class level; the red color indicates unsuccessful identification.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
JPY_662_sm_reshoot_supplementaryfig_s1.jpg1228KSupporting info item
JPY_662_sm_reshoot_supplementaryfig_s2.jpg1379KSupporting info item
JPY_662_sm_reshoot_supplementaryfig_s3.jpg1251KSupporting info item
JPY_662_sm_reshoot_supplementaryfig_s4.jpg1386KSupporting info item
JPY_662_sm_reshoot_supplementaryfig_s5.jpg1236KSupporting info item
JPY_662_sm_reshoot_supplementaryfig_s6.jpg1393KSupporting info item
JPY_662_sm_TableS1_supplementary.doc334KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.