An integrated analytical method for the separation, identification and quantification of carotenoids in Rhodobacter capsulatus

Abstract

The analysis and quantification of microbial carotenoids remain challenging due to the complexity of biological matrices and the high cost or limited availability of commercial standards. Herein, we present a cost-effective and reproducible workflow for the extraction, isolation, identification, and quantification of carotenoids in Rhodobacter capsulatus cultures. The method integrates light-protected pigment extraction, screening by thin-layer chromatography, and column chromatography for the isolation of individual carotenoids. Structural assignment is supported by UV–Vis spectroscopy, nuclear magnetic resonance (NMR), and LC-APCI-QTOF-HRMS exact-mass/isotopic-cluster evidence, while quantification is performed without commercial standards using the Lambert-Beer law with compound-specific molar extinction coefficients and HPLC-based calibration curves from internally isolated carotenoid-enriched fractions. The workflow showed strong calibration linearity (R² > 0.99), satisfactory reproducibility, with RSD values below 7 % for individual carotenoids, and high extraction efficiency recovering approximately 99 % of the total quantified carotenoids. The workflow facilitates comprehensive characterization of microbial carotenoid profiles, supporting both fundamental research and biotechnological applications. Future applications should evaluate its adaptation to other pigment-producing microorganisms and complex biological matrices.

Graphical abstract

Specifications table

Subject area	Chemistry
More specific subject area	Chromatographic and spectrophotometric methods for carotenoid analysis in microbial systems
Name of your method	Integrated method for the extraction, separation, identification and quantification of carotenoids in Rhodobacter capsulatus
Name and reference of original method	None
Resource availability	None

Background

The increasing global demand for sustainable and alternative protein sources has driven significant interest in microbial biomass as a novel food ingredient. In this context, purple phototrophic bacteria (PPB) such as Rhodobacter capsulatus have emerged as promising candidates due to their ability to convert organic substrates into valuable biomass while producing bioactive compounds, including carotenoids. These pigments are of particular interest because of their antioxidant properties and their role in human health [1]. Among the carotenoids produced by Rhodobacter capsulatus, neurosporene, spheroidene and spheroidenone are key components of the photosynthetic apparatus. Under anaerobic conditions, spheroidene is typically the dominant carotenoid, whereas in the presence of oxygen, the enzyme spheroidene monooxygenase catalyzes the incorporation of a keto group, leading to the formation of spheroidenone [2,3]. Reliable analytical methods for their identification and quantification are essential in both research and applied biotechnology contexts.

However, the analysis of microbial carotenoids presents several challenges. These compounds are typically found in low concentrations and complex biological matrices containing bacteriochlorophylls, lipids, and other pigments that can hinder their separation and detection. In addition, carotenoids are chemically sensitive to light, oxygen, and solvent conditions, which complicates their extraction and handling. Despite the growing interest in carotenoids, complete analytical workflows remain scarce. A Scopus® survey conducted over the last 20 years identified >20,700 publications related to carotenoid analysis, whereas only 20 articles specifically addressed carotenoid separation, identification, and quantification. This highlights the need for practical and accessible methodologies capable of reliably characterizing individual carotenoids in microbial systems. From an analytical perspective, absolute quantification commonly relies on pure reference standards, which are often expensive, limited in availability, or unavailable for specific microbial carotenoids [4]. Existing analytical approaches commonly involve chromatographic separation coupled with spectrophotometric or mass spectrometric detection [5]. Although these methods provide high sensitivity, the limited availability and high cost of carotenoid standards, often exceeding €1000 mg⁻¹, restrict their routine application (Table S1).

To address these limitations, this work presents an integrated analytical workflow for the extraction, isolation, identification, and quantification of the main carotenoids present in Rhodobacter capsulatus biomass. The methodology combines thin-layer chromatography for preliminary pigment screening, silica-based column chromatography for carotenoid enrichment, and complementary analytical techniques including UV–Vis spectroscopy, nuclear magnetic resonance (NMR), HPLC-DAD, and LC-APCI-QTOF-HRMS. The main novelty of the proposed approach is the generation of compound-specific HPLC-DAD calibration curves using internally isolated carotenoid-enriched fractions, thereby reducing dependence on costly or poorly available commercial standards. This practical and cost-effective workflow may facilitate carotenoid characterization, process optimization, and routine pigment monitoring in R. capsulatus, while providing a methodological basis for future adaptation to related pigment-producing microorganisms.

Method details

Microbial cultivation conditions

Rhodobacter capsulatus cultures were grown under controlled conditions to ensure reproducible carotenoid production. Cultivation was performed at 30 °C under continuous infrared illumination (850 nm) at an irradiance of 7.5 W m⁻², and continuous stirring at 500 rpm. Cultures were grown in RCV medium as described by Weaver et al. (1975) [6], but replacing the malate with acetate (20 mM) as soluble organic carbon source. Both strict anaerobic (N₂ atmosphere) and microaerobic (presence of O₂) conditions were evaluated. Cultures were incubated for approximately 100 h.

Biomass harvesting and pre-treatment

Biomass was first collected by centrifugation for 15 min at 4500 rpm and washed twice with phosphate-buffered saline (PBS) to remove residual medium components and impurities. All subsequent steps were performed under low-light or dark conditions to prevent carotenoid degradation.

Pigment extraction

Pigments were extracted by adding an organic solvent mixture (acetone:methanol, 7:2 v/v ratio) to the biomass pellet and stirring for approximately 1 h. Thereafter, samples were centrifuged to separate the liquid extract from the pellet followed by the removal of the supernatant obtained. This extraction step was performed twice, using 1 mL of acetone:methanol (7:2, v/v) in each cycle, and the two supernatants were combined.

Extract concentration and preparation

After extraction, the liquid pigment extracts were combined to obtain sufficient material for downstream analyses and concentrated to dryness under a continuous stream of nitrogen in dark conditions until complete solvent removal was achieved. Carotenoid degradation during extraction and concentration should be considered a critical aspect of the workflow, as these compounds are susceptible to photooxidation, thermal degradation, and trans–cis isomerization when exposed to light, oxygen, elevated temperature, or pro-oxidant conditions. In agreement with previous reports on β-carotene and related carotenoid degradation products, preventive measures during sample preparation, extraction, and analysis are required to avoid artificial oxidation and preserve the initial pigment profile [7]. To minimize these effects, all extraction and concentration steps were carried out using light-protective containers and under low-light or dark conditions. Solvent evaporation was performed under nitrogen to reduce oxygen exposure, and centrifugation steps were conducted at 4 °C using a refrigerated microcentrifuge. In addition, extracts were processed immediately after preparation, avoiding prolonged exposure to air, light, and room temperature.

Afterwards, the dried extract was re-dissolved in the minimum volume of hexane (≈2 mL) to obtain a highly concentrated pigment solution. This concentration step is critical to enable efficient separation and isolation of individual carotenoids in subsequent chromatographic analysis.

Thin-layer chromatography screening

Preliminary analysis of pigment composition was performed using thin-layer chromatography (TLC) as a rapid screening tool to (i) assess pigment diversity, (ii) evaluate the efficiency of different solvent systems, and (iii) guide the selection of appropriate conditions for subsequent column chromatography.

Neutral silica gel plates were used as the stationary phase to minimize carotenoid degradation. For the analysis, several mobile phase systems were initially tested to optimize pigment separation and resolution. Considering the highly hydrophobic nature of the target carotenoids and their interaction with the polar silica stationary phase, hexane was selected as the main non-polar eluent, while acetone was used as a mild polar modifier to promote controlled pigment migration. Based on these preliminary tests, the most effective mobile phase systems selected consisted of hexane:acetone mixtures, specifically 90–95 % hexane and 5–10 % acetone, which provided the best balance between resolution and pigment migration for carotenoid-enriched fractions.

Column chromatography for carotenoid isolation

Preparative separation of carotenoids was performed by column chromatography on silica gel. Silica gel was used as the stationary phase and packed initially with hexane. Pigment-concentrated samples obtained were loaded into the chromatographic column, and then a solvent gradient of hexane:acetone was added, with an initial v/v ratio of 95:5 and progressively changed to 90:10. This solvent gradient allowed improved separation of carotenoids based on polarity differences. The pigment fractions migrated through the column and were collected in different vials to carry out the subsequent analysis.

Carotenoid identification

The isolated pigment fractions corresponding to major carotenoids were analyzed using complementary techniques to perform their identification and quantification.

UV–Vis spectroscopy

A UV–Vis spectrophotometer (Thermo Scientific, Evolution 201) was used to acquire the absorption spectra of the samples and determine their absorption maxima and spectral fine structures.

Fig. 1 shows the absorption spectra of the major carotenoids present in the pigment fractions obtained following the previous methodology. These absorption spectra are consistent with neurosporene (λ_max ≈ 438 nm), spheroidene (λ_max ≈ 454 nm), and spheroidenone (λ_max ≈ 482 nm), in agreement with reported spectral data available in the LipidBank database [8] and literature on carotenoids from PPB [9,10].

Fig. 1.

Absorption spectra of carotenoids measured by UV–Vis spectrophotometry.

Nuclear magnetic resonance (NMR)

Liquid NMR analyses were performed to obtain structural confirmation of the carotenoids in the concentrated samples. ¹³C and ¹H NMR spectra were acquired using a VARIAN INFINITY 400 spectrometer operating at 79.4 MHz for ¹³C, with samples prepared in deuterated chloroform.

The NMR spectra of spheroidenone and spheroidene exhibit clear differences, primarily arising from the presence of a carbonyl group in spheroidenone. In the ¹³C NMR spectrum (Fig. 2(a)), spheroidenone shows a distinctive resonance at approximately 206 ppm, attributable to the ketone (C O) functionality. This signal is absent in spheroidene (Fig. 3(a)), whose ¹³C NMR spectrum is instead dominated by olefinic carbon signals in the ∼118–124 ppm range. Furthermore, several aliphatic carbon resonances in spheroidenone are shifted slightly downfield relative to those of spheroidene, particularly for carbons in proximity to the carbonyl group, reflecting its deshielding effect.

Fig. 2.

¹³C (a) and ¹H (b) NMR spectra of spheroidenone in CDCl₃.

Fig. 3.

¹³C (a) and ¹H (b) NMR spectra of spheroidene in CDCl₃.

In the ¹H NMR spectra (Fig. 2(b)), spheroidenone displays more downfield-shifted proton signals for hydrogen atoms located near the carbonyl group, whereas spheroidene (Fig. 3(b)), shows a more uniform pattern characteristic of a polyunsaturated hydrocarbon chain. Overall, these spectral differences highlight the electronic influence of the ketone group in spheroidenone compared to the purely hydrocarbon framework of spheroidene (Fig. 4). It should also be noted that the low concentration of these pigments results in reduced signal intensities in the NMR spectra.

Fig. 4.

Comparative ¹³C (a) and ¹H (b) NMR spectra of spheroidene (blue) and spheroidenone (red) recorded in CDCl₃, highlighting the presence of the carbonyl signal in spheroidenone and the overall spectral differences between both carotenoids.

HPLC-DAD analysis

Pigment-concentrated samples were also analyzed by HPLC-DAD to assess the chromatographic behavior of the target carotenoid peaks and to establish the methodology for their identification and quantification. The pigments contained in the samples were analyzed using an Agilent 1260 Infinity II HPLC system equipped with a diode array detector (DAD). A Gemini C18 reverse-phase column (150 mm × 3.0 mm, 5 μm particle size) maintained at 20 °C was used with a mobile phase of acetonitrile:methanol (70:30, v/v) at a flow rate of 0.5 mL min⁻¹. The injection volume was 20 μL and the total run time was 17 min. DAD spectra were recorded over the range of 260–600 nm, and chromatographic peak areas were integrated at the compound-specific absorption maxima used for quantification: neurosporene (438 nm), spheroidene (454 nm), and spheroidenone (482 nm).

The acetonitrile:methanol mixture was selected as a fully organic mobile phase suitable for reversed-phase C18 separation of hydrophobic carotenoids. Acetonitrile provides strong elution strength and low viscosity, whereas methanol contributes to chromatographic selectivity [11]. Different acetonitrile:methanol proportions, including 90:10, 80:20, and 70:30 (v/v), were preliminarily evaluated under otherwise identical chromatographic conditions. Among the tested compositions, the 70:30 (v/v) mixture provided the best compromise between retention, peak resolution, and chromatographic selectivity for the target carotenoids. Under these conditions, spheroidenone eluted first, followed by spheroidene and neurosporene. The absence of water in the mobile phase also minimized solubility issues during routine HPLC-DAD analysis.

The HPLC-DAD analysis showed a single main target peak for each isolated carotenoid fraction at the corresponding retention time: spheroidenone (RT = 8.0 min), spheroidene (RT = 10.7 min), and neurosporene (RT = 14.9 min) (Fig. 5). Furthermore, 3D chromatographic profiles were obtained using Agilent OpenLab CDS software. Together with the compound-specific DAD absorption, these results support the chromatographic suitability of the isolated fractions for target peak-area-based calibration. The signal observed at 1.8 min was attributed to the injection solvent (acetone:methanol, 7:2 v/v ratio).

Fig. 5.

Integrated HPLC-DAD chromatograms and chemical structures (a) and 3D chromatographic profiles (b) of the identified carotenoids.

LC-APCI-QTOF-HRMS analysis

To provide additional orthogonal evidence for carotenoid identification, the carotenoid-enriched fractions were analyzed by LC-APCI-QTOF-HRMS using an Impact Q-TOF mass spectrometer. Liquid chromatography enabled chromatographic separation of the carotenoid fractions before mass spectrometric detection, while atmospheric pressure chemical ionization (APCI) was selected because it is suitable for relatively hydrophobic and poorly polar compounds such as carotenoids [12]. The QTOF-HRMS system enabled high-resolution exact-mass measurement and comparison of the experimental isotopic pattern with the theoretical distribution expected for each molecular formula.

The experimental spectra were compared with the theoretical isotopic patterns calculated for the protonated molecules [M+H]⁺ of neurosporene, spheroidene, and spheroidenone. Theoretical exact masses and isotopic patterns were calculated from the elemental composition of each carotenoid using the Compass IsotopePattern tool. For the three carotenoid-enriched fractions, the monoisotopic exact masses and the main isotopic signals (M, M + 1, and M + 2) were consistent with the theoretical clusters, with monoisotopic mass errors not exceeding 3 ppm (Table 1). These results provide additional evidence supporting the assignment of the three target carotenoids, complementing the UV–Vis, NMR, and HPLC-DAD data. Representative LC-APCI-QTOF-HRMS chromatographic profiles and experimental HRMS spectra for the three carotenoid-enriched fractions are provided in the Supplementary Information (Figure S1).

Table 1.

LC-APCI-QTOF-HRMS exact-mass and isotopic-cluster data supporting the identification of neurosporene, spheroidene, and spheroidenone.

Carotenoid	Molecular formula	Ion	Theoretical isotopic cluster m/z M / M + 1 / M + 2	Experimental isotopic cluster m/z M / M + 1 / M + 2	Monoisotopic error (ppm)
Neurosporene	C₄₀H₅₈	[M+H]⁺	539.4611/540.4654/541.4679	539.4617/540.4652/541.4682	+1.1
Spheroidene	C₄₁H₆₀O	[M+H]⁺	569.4717/570.4751/571.4784	569.4725/570.4759/571.4788	+1.4
Spheroidenone	C₄₁H₅₈O₂	[M+H]⁺	583.4510/584.4544/585.4577	583.4527/584.4561/585.4592	+3.0

Quantification of carotenoids

Spectrophotometric quantification

Quantification was performed using the Lambert-Beer law (Eq. (1)), applying solvent-specific molar extinction coefficients. Spectra of isolated carotenoid extracts were acquired in the appropriate solvent for each compound, ensuring consistency with the extinction coefficients used for quantification. This alignment is critical, as molar absorptivity values are solvent-dependent and directly influence concentration accuracy. The concentration of each carotenoid was calculated from its λ_max using literature-reported extinction coefficients specific to each solvent system [9,10] (Table 2).

$$A=\epsilon ·c·l$$ (1)

Table 2.

Spectroscopic parameters, molar mass, and solvent conditions used for carotenoid quantification.

Carotenoid	ε (mM⁻¹·cm⁻¹)	Molar mass (g·mol⁻¹)	λmax (nm)	Solvent
Neurosporene	159.4	538.9	438	n-Hexane
Spheroidene	173.6	568.9	454	n-Hexane
Spheroidenone	122	582.9	482	Acetone:methanol (7:2, v/v)

Where A is the absorbance, ε is the molar extinction coefficient (mM⁻¹·cm⁻¹), c is the carotenoid concentration (mM), and l is the optical path length (cm). Accordingly, carotenoid concentrations were first calculated in mM and then converted to mg·L⁻¹ using the molar mass of each compound.

Because the initial concentration of each carotenoid fraction was assigned using literature-reported extinction coefficients, this step represents a potential source of systematic uncertainty. Extinction coefficients are compound and solvent dependent and may also be influenced by pigment purity, isomeric composition, wavelength accuracy, and instrumental conditions. Previous work using spheroidene and related bacterial carotenoids has suggested that these coefficients may introduce an uncertainty of approximately ±5 % [13]. Therefore, any uncertainty in ε is directly propagated to the calculated concentration of the isolated fractions and subsequently to the HPLC calibration curves generated from them. In this work, this uncertainty was minimized by using compound-specific extinction coefficients reported for the same solvent systems, measuring absorbance at the corresponding λ_max, and using internally characterized carotenoid-enriched calibration materials. Nevertheless, the resulting concentrations should be interpreted as values traceable to literature extinction coefficients rather than to certified commercial reference standards.

Calibration curve generation

For routine HPLC quantification of carotenoids, calibration curves were constructed using the concentrated samples of known concentration, obtained following the methodology described previously.

Before injection into the HPLC system, the isolated carotenoid-enriched fractions were dried and re-dissolved in a solvent compatible with the chromatographic protocol, maintaining the same total volume. In this study, neurosporene and spheroidene, originally dissolved in hexane, were dried and subsequently re-dissolved in acetone:methanol (7:2 v/v), whereas spheroidenone could be analyzed directly without further manipulation.

For the HPLC analysis, a dilution series of each carotenoid was prepared, and samples were injected in triplicate, recording the corresponding peak areas obtained. Thereafter, calibration curves were generated by plotting HPLC peak areas against the known carotenoid concentration (mg·L⁻¹) (Figure S2). The raw calibration datasets used for curve construction are provided in Tables S2–S4. These calibration curves allow direct conversion of chromatographic signals from unknown samples into routine carotenoid concentrations. The isolated fractions were used as internally purified calibration materials rather than as certified analytical standards. Calibration was based on the integrated area of the target carotenoid peak at the compound-specific wavelength. Therefore, impurities or background signals outside the target retention time were not included in the calibration response.

Routine sample quantification

Once calibration curves have been established, routine samples can be quantified by injecting the final pigment extracts into the HPLC-DAD system and converting target peak areas into carotenoid concentrations using the corresponding compound-specific calibration equations.

Critical considerations and practical notes

To correctly apply the methodology described in this work, several practical aspects should be considered. As discussed in Section 4, carotenoids are sensitive to light, oxygen, temperature, and prolonged handling times, therefore, all extraction, concentration, and chromatographic steps should be performed under low-light conditions. Fresh pigment extracts should be analyzed immediately after preparation, and solvent evaporation should be performed under nitrogen to reduce oxidative degradation. Temperature control is recommended during sample handling, and centrifugation at 4 °C together with rapid extract processing helps reduce thermal degradation and preserve the original pigment profile. Finally, neutral silica gel should be used for TLC and column chromatography, since acidic or highly reactive chromatographic supports may promote pigment degradation or isomerization.

Method validation

Linearity and instrumental precision

Linearity and instrumental precision were evaluated from the calibration curves generated as described in Section 8.2. Table 3 summarizes the calibration range, regression equation, coefficient of determination (R²), instrumental precision, limit of detection (LOD), and limit of quantification (LOQ) obtained for neurosporene, spheroidene, and spheroidenone. All target carotenoids showed strong linearity within the evaluated concentration ranges, with R² values higher than 0.99.

Table 3.

Calibration parameters, linearity, instrumental precision, LOD and LOQ for carotenoid quantification by HPLC-DAD.

Carotenoid	Range (mg L⁻¹)	Equation	R²	RSD (%)	LOD (mg L⁻¹)	LOQ (mg L⁻¹)	n
Neurosporene	0.21–2.64	y = 537.44 x − 0.5986	0.9988	0.8–5.0	0.11	0.34	5
Spheroidene	0.77–12.25	y = 801.11 x + 5.6929	0.9997	0.1–0.8	0.27	0.82	5
Spheroidenone	1.36–14.13	y = 499.28 x + 28.758	0.9990	0.1–2.3	0.54	1.63	5

Instrumental precision was evaluated by calculating the relative standard deviation (RSD, %) of replicate injections at each calibration level. The HPLC-DAD system showed low instrumental variability across the calibration range, with RSD values remaining around or below 5 %, demonstrating consistent repeatability of the chromatographic response.

Limits of detection (LOD) and quantification (LOQ) were estimated from the calibration curves using Eqs. (2) and 3:

$$\text{LOD}=3.3·\sigma /\mathrm{S}$$ (2)

$$\text{LOQ}=10·\sigma /\mathrm{S}$$ (3)

where σ is the residual standard deviation of the calibration curve and S is the slope.

The residual analysis of the calibration models, calculated as the difference between the experimental mean HPLC peak area and the area predicted by each calibration equation, is provided in the Supplementary Information (Table S5).

Method reproducibility

The reproducibility of the analytical workflow was assessed using six independent replicate samples for each culture condition (N₂ and O₂). Each replicate was subjected to the complete workflow, including pigment extraction, chromatographic analysis, and calibration-based quantification.

For each carotenoid, concentration values in the final carotenoid extract analyzed by HPLC-DAD (mg L⁻¹ extract) were obtained and used to calculate mean values, standard deviation (SD), and relative standard deviation (RSD, %). The RSD values reported in Table 4 therefore reflect the variability associated with the complete analytical procedure, including sample heterogeneity, extraction, sample handling, and instrumental quantification.

Table 4.

Quantification reproducibility of carotenoids in final extracts obtained from R. capsulatus cultures under N₂ and O₂ conditions using six replicate samples per condition.

Condition	Compound	Mean (mg L⁻¹ extract)	SD	RSD (%)
N₂	Neurosporene	<LOD (0.08)	0.01	6.9
	Spheroidene	3.65	0.11	3.1
	Spheroidenone	<LOD (0.53)	0.03	5.7
	TOTAL	4.26	0.14	3.2
O₂	Neurosporene	0.36	0.01	2.2
	Spheroidene	1.60	0.09	5.9
	Spheroidenone	4.06	0.11	2.7
	TOTAL	6.01	0.09	1.4

Despite the complexity of the microbial matrix and the multi-step nature of the workflow, the method showed satisfactory reproducibility, with individual carotenoid RSD values remaining below 7 % under both culture conditions. Under N₂ conditions, spheroidene was the dominant carotenoid, whereas under O₂ conditions, spheroidenone was the major compound detected. These results confirm that the method can be reproducibly applied to samples with different carotenoid distributions.

In addition to individual carotenoid assessment, total carotenoid concentration was estimated as the sum of neurosporene, spheroidene, and spheroidenone concentrations, including low-level estimated signals below the validated LOQ where applicable. Total quantified carotenoids also showed low variability, with RSD values close to 3 % under N₂ and below 2 % under O₂ conditions. These results support the robustness of the workflow for routine quantification of the major carotenoids in R. capsulatus. To ensure full analytical traceability, the supplementary information provides the corresponding HPLC peak areas and calibration-based concentration values for the replicate samples analyzed under N₂ and O₂ conditions (Table S6).

Values in parentheses for compounds below the LOD or LOQ correspond to calibration-based estimated concentrations calculated from the target HPLC-DAD peak area. These values are reported for traceability and should be interpreted as semi-quantitative estimates.

Extraction efficiency

The efficiency of the pigment extraction procedure was evaluated by analyzing three consecutive extraction cycles separately. For each culture condition (N₂ and O₂), three replicate pellets were subjected to sequential extraction with acetone:methanol (7:2, v/v). The extracts obtained after the first, second, and third extraction cycles were collected independently and analyzed by HPLC-DAD.

Each extraction cycle was performed using 1 mL of solvent. Since the same extraction volume was applied to extract 1 (E1), extract 2 (E2), and extract 3 (E3), the relative contribution of each extraction step was calculated directly from the corresponding target peak areas. For each carotenoid, the total chromatographic response was defined as the sum of the peak areas obtained across the three consecutive extraction cycles. The efficiency of the routine extraction protocol was calculated as the cumulative percentage accounted for by the first two extraction cycles, whereas the third extraction was considered a residual extraction.

The results are summarized in Table 5. The first extraction cycle recovered the largest fraction of the extracted carotenoids in all cases, while the second extraction contributed a smaller but relevant fraction. The third extraction represented only a minor residual fraction. Considering individual carotenoids, the first two extraction cycles recovered between 96.95 and 100.00 % of the total amount recovered after three extractions. When the combined HPLC-DAD response of the three target carotenoids was considered, the first two extraction cycles accounted for 98.97 ± 0.65 % under N₂ conditions and 98.94 ± 0.07 % under O₂ conditions. These results indicate that two sequential extraction cycles are sufficient to recover most of the extractable carotenoid content under the conditions tested.

Table 5.

Sequential extraction efficiency of carotenoids from R. capsulatus extracts.

Condition	Compound	E1 (%)	E2 (%)	E3 (%)	E1+E2 efficiency (%)
N₂	Neurosporene	84.02 ± 8.77	15.98 ± 8.77	0.00 ± 0.00	100.00 ± 0.00
N₂	Spheroidene	87.41 ± 6.57	11.88 ± 6.23	0.72 ± 0.34	99.28 ± 0.34
N₂	Spheroidenone	75.30 ± 15.66	21.64 ± 13.18	3.05 ± 2.48	96.95 ± 2.48
N₂	Total	85.67 ± 7.99	13.30 ± 7.34	1.03 ± 0.65	98.97 ± 0.65
O₂	Neurosporene	90.41 ± 1.09	9.59 ± 1.09	0.00 ± 0.00	100.00 ± 0.00
O₂	Spheroidene	81.74 ± 4.43	16.84 ± 4.12	1.42 ± 0.37	98.58 ± 0.37
O₂	Spheroidenone	79.75 ± 1.93	19.26 ± 1.74	0.99 ± 0.19	99.01 ± 0.19
O₂	Total	80.78 ± 0.30	18.16 ± 0.23	1.06 ± 0.07	98.94 ± 0.07

Values are expressed as mean ± SD (n = 3). E1, E2, and E3 correspond to the first, second, and third extraction cycles, respectively. E1 + E2 represents the cumulative efficiency of the routine extraction protocol, whereas E3 was included as a residual extraction step. Total values were calculated from the summed HPLC-DAD target peak areas of neurosporene, spheroidene, and spheroidenone obtained in each extraction cycle.

For neurosporene, no detectable peak was observed in the third extraction cycle under either culture condition. These values were treated as zero for the calculation of extraction percentages. Representative UV–Vis absorption spectra of the sequential extracts obtained after the first, second, and third extraction cycles under N₂ and O₂ conditions are provided in the Supplementary Information (Figure S3). These spectra qualitatively illustrate the progressive decrease in absorbance across successive extraction cycles and support the residual character of the third extraction. However, extraction efficiency was calculated from the relative HPLC-DAD target peak areas rather than from UV–Vis absorbance.

Apparent recovery assay

To further evaluate the performance of the proposed extraction and quantification workflow, an apparent recovery assay was performed using spheroidene as a representative target carotenoid. Spheroidene was selected because it was the main carotenoid detected under anaerobic conditions and one of the major compounds quantified in the proposed method.

For this assay, Rhodobacter capsulatus biomass pellets obtained from anaerobic cultures were divided into non-spiked controls and spiked samples. A known volume of a spheroidene-enriched fraction, containing a nominal spheroidene amount of 6.00 µg, was added directly to each spiked biomass pellet before extraction. The solvent was subsequently removed under a gentle nitrogen stream in dark conditions, leaving the added pigment deposited onto the biomass. The pellets were then processed using the routine two-step extraction protocol with acetone:methanol (7:2, v/v), resulting in a final combined extract volume of 2.00 mL. Therefore, the added spheroidene amount corresponded to a theoretical concentration of 3.00 mg L⁻¹ in the final extract. Non-spiked control pellets were processed in parallel using the same procedure but without spheroidene addition.

Recovery was calculated by comparing the increase in measured spheroidene concentration in spiked samples with the theoretical concentration added, after subtracting the endogenous spheroidene concentration measured in the non-spiked controls, according to Eq. (4):

$$\text{Apparent}\text{recovery}(\%)=\left[\left(C_{\text{spiked}}-C_{\text{control}}\right)/C_{\text{added}}\right]\times 100$$ (4)

where C_spiked is the mean spheroidene concentration measured in the spiked samples, C_control is the mean endogenous spheroidene concentration measured in the non-spiked controls, and C_added is the theoretical spheroidene concentration in the final extract corresponding to the nominal amount added to the biomass pellet.

The mean spheroidene concentration measured in non-spiked controls was 5.17 ± 0.30 mg L⁻¹, whereas spiked samples showed a mean concentration of 8.11 ± 0.20 mg L⁻¹. Considering a theoretical added concentration of 3.00 mg L⁻¹, the apparent recovery of spheroidene was 98.0 % (Table 6). These results support the suitability of the extraction and HPLC-DAD quantification workflow for the recovery and routine quantification of the dominant carotenoid under anaerobic conditions. However, this assay should be interpreted as a representative apparent recovery assessment rather than a full compound-specific recovery validation for all target carotenoids.

Table 6.

Apparent recovery of spheroidene in R. capsulatus biomass pellets under anaerobic conditions.

Carotenoid	Condition	Ccontrol (mg L⁻¹)	Cadded (mg L⁻¹)	Cspiked (mg L⁻¹)	Recovery (%)
Spheroidene	N₂	5.17 ± 0.30	3.00	8.11 ± 0.20	98.0

Conclusions

Analytical performance and methodological advantages

The proposed workflow provides an integrated strategy for carotenoid analysis in Rhodobacter capsulatus, combining pigment extraction, chromatographic enrichment, complementary compound identification, and HPLC-DAD quantification within a single analytical framework. Its main methodological advantage is the generation of compound-specific calibration curves using internally isolated carotenoid fractions, thereby reducing dependence on costly or poorly available commercial standards.

The workflow also showed suitable analytical performance for routine application, including strong calibration linearity, low instrumental variability, high extraction efficiency, and satisfactory apparent recovery for spheroidene. These results support its use for reproducible quantification and comparative monitoring of the major spheroidene-pathway carotenoids.

A further advantage is the complementary and modular use of different analytical platforms. UV–Vis spectroscopy provides rapid spectral screening and compound-specific absorption information, whereas NMR and LC-APCI-QTOF-HRMS provide additional structural and exact-mass evidence during compound identification. Once carotenoid identity, retention behavior, and calibration have been established, routine analysis can be performed primarily by HPLC-DAD, reducing the need to apply the complete analytical workflow to every sample.

This modular approach makes the method particularly suitable for repeated monitoring of carotenoid profiles during cultivation, process optimization, and experiments involving changes in pigment composition. Rather than aiming to generate certified analytical standards, the workflow provides a practical internal calibration strategy for routine and comparative analysis of neurosporene, spheroidene, and spheroidenone in R. capsulatus cultures.

Limitations

Despite its analytical performance, several limitations of the proposed workflow should be considered. First, the complete identification strategy relies on access to complementary analytical platforms. HPLC-DAD is required for chromatographic separation and routine quantification, whereas NMR and LC-APCI-QTOF-HRMS provide additional structural and exact-mass evidence. Access to these specialized instruments, trained personnel, and sufficient amounts of isolated pigment may therefore limit the implementation of the complete workflow in laboratories with restricted analytical infrastructure. Nevertheless, once compound identity and chromatographic behavior have been established, routine monitoring can be performed primarily by HPLC-DAD.

The workflow was specifically optimized for the main spheroidene-pathway carotenoids produced by Rhodobacter capsulatus. Its application to other microorganisms or more complex microbial communities may require modification of the extraction conditions, chromatographic separation, detection wavelengths, and calibration strategy. Differences in pigment polarity and spectral properties may also lead to partial chromatographic or spectroscopic overlap.

A further limitation is the requirement to isolate individual carotenoids before constructing the internal calibration curves. This step may be particularly challenging for low-abundance compounds, for which insufficient material may be recovered for complete structural characterization and reliable calibration. Preparative column chromatography also increases sample handling and analysis time and requires strict protection from light and oxygen throughout the isolation procedure.

Finally, the use of internally isolated carotenoid fractions reduces dependence on expensive or poorly available commercial standards but does not provide the same metrological traceability as certified reference materials. Quantitative results may be influenced by the composition and isomeric profile of the isolated fractions, the literature-derived extinction coefficients used to assign their initial concentrations, and potential matrix effects during extraction and HPLC-DAD analysis. Accordingly, the method is primarily intended for routine and comparative quantification within the described analytical framework.

Future perspectives

The workflow described in this study was optimized and validated for spheroidene-pathway carotenoids in Rhodobacter capsulatus. Nevertheless, its extension to other pigment-producing microorganisms, particularly purple phototrophic bacteria (PPB), represents a relevant future application. Many of these microorganisms synthesize structurally related acyclic C40 carotenoids with comparable hydrophobic character, although their degree of conjugation, oxygenation/methoxylation pattern, polarity, absorption maxima, and chromatographic behavior may differ substantially.

A representative example is Rhodospirillum rubrum, which accumulates spirilloxanthin-type carotenoids. Spirilloxanthin and anhydrorhodovibrin have been described as methoxylated, highly conjugated C40 carotenoids, and their HPLC separation has been specifically addressed through optimization of solvent conditions [14]. This supports the idea that organic solvent-based extraction and chromatographic workflows can be adapted to carotenoids beyond the spheroidene pathway, while also highlighting that method transfer requires compound-specific optimization.

Therefore, future applications of the proposed workflow should evaluate its adaptation to other purple phototrophic bacteria and carotenoid families. Such adaptation would require re-evaluation of the extraction solvent system, TLC and column chromatography conditions, HPLC-DAD mobile phase, detection wavelengths, and extinction coefficients used for quantification. Accordingly, the present method should be regarded as a flexible analytical framework rather than as a universal protocol directly transferable to all carotenoid-producing microorganisms.

Related research article

None.

For a published article

None.

Ethics statements

Not applicable.

CRediT authorship contribution statement

Javier Llama: Conceptualization, Investigation, Writing – original draft. Sara Jerez: Conceptualization, Investigation, Writing – original draft. María Ventura: Investigation, Methodology, Validation, Writing – review & editing. Amanda Prado: Investigation, Methodology, Validation, Writing – review & editing. Daniel Puyol: Funding acquisition, Supervision, Writing – review & editing. Juan Antonio Melero: Funding acquisition, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.