A critical review of analytical methods used for the chemical characterisation and quantification of phlorotannin compounds in brown seaweeds

Abstract Introduction Phlorotannins, the phenolic compounds found in brown seaweeds, are a unique and diverse class of compounds showing a huge potential for food and pharmaceutical applications. Objective This review will give an account of the colorimetric assays used and a discussion of their quantitative and qualitative analytical shortcomings. It will also discuss other more complex and modern analytical chemistry methods that are currently being developed to study phlorotannins. The purpose of this review is to increase awareness of these bioactive compounds and promote further development of robust analytical methods for use in biology, food science, pharmacology and biomedical and cosmeceutical sciences. Results Whilst the biological activity and huge commercial potential of the phlorotannins has been widely reported throughout the literature, the chemical structures and reactivity of these compounds is still not well understood. The phlorotannin content of seaweed is usually characterised using colorimetric assays. However, although these methods give a reasonable overall estimation of the total phenolic content, they lack precision and specificity. Conclusion This review highlights the strengths and weaknesses of commonly used colorimetric assays. Novel techniques are highlighted using more selective chemistry to identify this class of compounds.

unsaturated fatty acids, 5-7 alginate 8-10 and biopolymers. 11 Seaweed supplemented animal feeds offer nutritionally rich renewable feedstocks with several positive attributes to livestock health. 5,10,[12][13][14] These attributes include improved gut health and reduced faecal shedding. 13,15 It has also been reported in the literature that supplementing seaweed into the diet of pigs can decrease the need for antibiotics, which could slow or even avert the antibiotic resistance crisis the livestock industry is currently facing. 16,17 These antimicrobial and antibacterial effects have been reportedly due to the complex matrix of polyphenolic structures in the plant, collectively called phlorotannins. [18][19][20] Brown seaweeds have also been shown to reduce methane emissions when fed to ruminants, which is of major importance in the context of the increasingly restricted emission targets that impinge on this area of intensive farming. 5 Phlorotannins have been shown to be immensely valuable to the food and cosmetic industries due to their antioxidant and anti-inflammatory properties. 21,22 Despite this, little is still understood about the characterisation and chemistry of these molecules. In order to gain a better understanding of the industrial value of phlorotannins, robust analytical methods are required to map the structure and size, of the compounds and their chemical linkages. Seaweed farming is of significant economic importance to immediate coastal areas 1 and can generate higher revenue streams through the processing of high value products, such as phlorotannins.
Phlorotannins display very similar characteristics to tannins produced by terrestrial plants but are structurally very different. 4 Phlorotannins are polymeric structures of the monomer phloroglucinol (1,3,5-hydroxybenzene) that are mainly linked through aryl-aryl C-C bonds or aryl-ether C-O bonds. The naming of phlorotannins is systematic based on the types of linkages between the aromatic groups, fucols only consist of aryl-aryl linkages ( Figure 1) whereas phloroethols are exclusively ether linkages through the phenolic oxygen. 23 However, it is also possible to have both types of linkage in a single phlorotannin compound, which is named as a phlorofucol. 24 Due to the variety of linkages and the array of molecular sizes of the phlorotannins, the number of possible structures is very high and as the molecular weight increases, the chemical complexity of the structure also increases. The molecular weight of this class of compounds can range from 126 to >10000 Da. 23,25 Phlorotannins have only been reported to be produced by brown seaweeds and are biosynthesised by the acetate malonate pathway. 26,27 They are very hydrophilic due to the presence of many phenolic OHs in their structure. This allows easy absorption of phlorotannins into biological systems when digested. Phlorotannins are most highly concentrated in the epidermal cortex of brown seaweed, however they are also found to be bound into the cell wall of the marine macroalgae. 27 Phlorotannins are reported to contribute to the defence of the plant by acting as a herbivore deterrent, possibly by acting as an appetite suppressant. 28 Phlorotannin concentrations vary from 0.5% to 20% of their dry weight, which can fluctuate with respect to season (e.g. changes in light exposure), environment (e.g. nutrient availability in the surrounding water) and also between species. 29 It is noteworthy that there are some significant discrepancies between results in the literature and their methods of calculation. This is of particular concern from an analytical chemistry perspective, as yields quoted can mislead industry who are interested in commercialising particular phytochemicals.
Tannins from terrestrial plants are considered to have many health benefits including antioxidant properties, which have been claimed to help prevent degenerative diseases such as cancer and Alzheimer's. 19 There have been many reviews addressing the biological activity of phlorotannins 21,[30][31][32][33] and their effects in food digestion for animals 12 and humans has been a prevalent topic in recent literature.

FIGURE 1 Phlorotannin structures
Phlorotannins have also been suggested to prevent the degradation of hyaluronic acid through the inhibition of HAase, which is linked to skin ageing, inflammation and also the migration of cancer cells. 26 Antioxidants are sacrificial reducing agents, which minimise the effects of oxidation on a substrate, mechanism shown in Figure 3. Food rancidity is often caused by oxidation and hence the use of antioxidants can increase the shelf life of food. There are many diseases also associated with oxidative stress, and a diet high in antioxidants is thought to minimise the onset of these conditions. 28 The oxidative stress in a biological system is caused by the presence of reactive oxygen species (ROS). These ROS occur naturally from the metabolism of cellular organisms or can be due to environmental factors such as smoke and pollution. If ROS are not quenched by the sacrificial reduction of an antioxidant they will react with carbohydrates, proteins, lipids or nucleic acids in biological systems and can disrupt their functionality. 28 The phenolic content of marine (seaweed) and terrestrial plants is of interest across many fields studying plant matrices, due to their roles in anti-oxidative reactions, astringency, bitterness, colour and browning reactions in food and formulation. 34

| TOTAL PHENOLIC CONTENT ASSAYS -QUANTITATIVE TECHNIQUES
Assays commonly used to study the polyphenolic content in terrestrial plants have also been extensively applied to study seaweeds. 8,[35][36][37][38] There are a number of assays which have been used to assess the content and activity of the polyphenolic phlorotannin compounds in seaweeds. The most commonly used colorimetric assays for the phlorotannin content are Folin-Ciocalteu (F-C) and 2,4dimethoxybenzaldehyde (DMBA) assays (Table 2). These two assays vary slightly depending on both the way they interact with the target compounds and their sensitivity.
The F-C assay is the most common assay used to quantify the phenolic content in both terrestrial plants and seaweeds. 8,35,38,[39][40][41] The F-C reagent is made up of a mixture of tungsten and molybdate. 42 The F-C assay relies on the transfer of electrons from phenolic compounds to phosphomolybdic/phosphotungstic acid complexes in alkaline conditions (Figure 2). 43 The transfer of these electrons facilitates a colour change, which can then be detected at 760 nm in the visible spectrum. 44 The blue colour which occurs upon reduction of the F-C reagent is thought to be due to a coordinated molybdenum(V) species, although full characterisation of the structure is still unknown. 42 Using the Beer-Lambert law the absorbance of the blue wavelength can then be calibrated using a standard compound. The phenolic content is calculated as equivalents of the standard. 43 This method lacks sensitivity and its accuracy is questionable due to interference from other substrates (e.g. soluble sugars and proteins) in the plant matrix that can react with the F-C reagents, causing artificially high values. 45 Phenols, proteins, and thiols were all found to cause a reaction with the F-C reagent in a study testing 80 different compounds. 42 Many vitamins were also found to cause a response along with the inorganic ions Fe 2+ , Mn 2+ , I − , and SO 3+ . 42 Therefore, the F-C reaction is not sensitive to phenols but a measurement of anything in the matrix that is able to reduce, and therefore be oxidised itself. There is an assumption that the majority of reducing power in a plant matrix is due to phenolics and therefore the F-C assay is used to give an approximation of the phenolic content.
DMBA reacts specifically with 1,3-and 1,3,5-trihydroxybenzenes in a similar reaction to the vanillin assay to flavonoid compounds. 37 This facilitates a colour change, which can be observed in the visible spectrum at 510 nm 37 and quantified from a calibration curve of a standard. This is therefore more selective for the phenolic structures present in seaweed, as the specific functionality is required to react with the reagent, and so is less prone to overestimated content than the F-C method. However, upon analysis of phlorotannins, this could possibly lead to an underestimation of their concentration as some phlorotannins are branched and can also have aryl linkages or fuhalols which contain additional hydroxyl groups in their structure in the 2, 4 or 6 positions. The mechanism of this reaction is not fully understood but DMBA is used for phlorotannin analysis over vanillin due to a stronger colour being produced. The reaction with the DMBA reagent must occur between the hydroxyl groups through the 2, 4 or 6 positions, therefore if these positions are substituted there will be no colour change and these phenolics will not be detected. This method must therefore be exercised with caution because some species could have more of these branches or linkages than others and therefore result in lower phenolic content reading than is actually the case.
Steric hindrances could also cause lower reaction rates in larger phlorotannins with a larger degree of polymerisation. Table 2 lists the various extraction methods and conditions used to measure the phenolic contents in seaweed. Table 2 highlights the variation between the protocols used in the literature, which then makes it difficult to directly compare between studies. Pre-treatment of seaweed prior to extraction can also make a large difference on the results, and hence standardisation of the pre-treatments when analysing phenolic content is also needed. Table 2 shows some studies use hexane as a pre-treatment to remove lipophilic substances prior to F-C analysis. This could help to remove fatty acids and other substances such as pigments that interfere with colorimetric assays. However, current protocols are not standardised, thus being another potential cause of discrepancies in results. Dichloromethane is also occasionally used to remove lipids from seaweeds; however, this process involves greater risk as dichloromethane is much more polar (solvent polarity index, 3.1) than hexane (solvent polarity index, 0.1) and hence could remove some compounds of interest.
There are a number of different parameters which need to be considered when analysing the phenolic content in the seaweeds, from collection through to measurement. Some studies fail to note the col- to be the best extraction method for highest total phenolic content (TPC). 46 This paper also assessed the optimum temperature for drying, extraction and pre-treatment. They suggest the storage of dried seaweed at 4°C prior to use, a pre-treatment with hexane (solid/liquid ratio of 5:1, w/v) and an extraction temperature of 55°C. Leyton However, these conditions could vary from laboratory to laboratory, based on the room temperature in that part of the world. It is also known that the phlorotannins in brown seaweeds readily undergo oxidation and hence overnight extraction could lead to depleted results. [48][49][50][51] It is impossible to get a direct measurement of the total percentage weight of the phenolics, using these colorimetric assays. This is due to the vast quantities of different structures, hence a calibration curve of a known phenolic is created and used to express the phenolic quantities as an equivalent. There is a need to express the phenolic content with the same phenolic equivalent, so the results can be fit for comparison. The majority of the literature measuring the phenolic content of seaweed is currently using phloroglucinol (Table 2), which is logical as this is the monomer unit of the phlorotannin polymer being measured and hence would be most likely to show similar results.
Phloroglucinol is the most widely used equivalent method to use when comparing between seaweed species, however, the majority of studies analysing terrestrial plants use gallic acid equivalents. Hence, if a comparison of the phenolic content of seaweed to terrestrial plants is required, a calibration of both phloroglucinol and gallic acid could be undertaken in order to get gallic acid equivalents and phloroglucinol equivalents.
Koivikko et al. 38 studied different methods in their extraction procedure to combat the rapid oxidation of the phlorotannins. In this study, a range of conditions were tested, such as, adding ascorbic acid (an antioxidant), lowering the pH, and reducing the temperature. The addition of ascorbic acid is used because of the ease of oxidation of the phlorotannins, which could be protected by the antioxidant nature of ascorbic acid, 52 it is thought there will be sacrificial oxidation to inhibit the oxidation of the phorotannins before analysis. Samples were then analysed by the F-C assay and both normal and reversed phase high-performance liquid chromatography (HPLC). Whilst the addition of ascorbic acid seemed beneficial when analysing by HPLC (more peaks observed due to less oxidation), it can cause interferences when using colorimetric assays, e.g. F-C assay, as it would react with the F-C reagent. Hence in order to avoid getting a more concentrated reading from the addition of ascorbic acid in the extraction, a reading of ascorbic acid only (at the same concentration that was added to the extraction) with the F-C dye needs to be taken from the final calculation.  This has been carried out by Parys et al., who reported that NMR spectroscopy recorded higher TPC levels than the F-C assay. 36 However, both methods did follow the same trend in seasonal variations.
The internal standard used was trimesic acid in 0.8 mL deuterated methanol and 0.2 mL deuterium oxide. A calibration curve was produced using phloroglucinol but it is important to note that when using NMR spectroscopy, the direct molar concentration of phlorotannin phenols are measured and it is not a calculated phloroglucinol equivalent. Since trimesic acid is used as the internal standard and not phloroglucinol, a comparison cannot be accurately made between this method and the other F-C assays.
High-resolution magic angle spinning (HR-MAS) has also be used to quantitatively measure the phlorotannins in the brown algae; Cystoseira tamariscifolia. 54 In this study solid-state NMR was used to observe the presence of phloroglucinol in a solid sample of seaweed.
The phloroglucinol singlet at 6.02 ppm was used to indicate presence within a sample with a relative intensity of three aromatic protons.

| LIQUID CHROMATOGRAPHY OF PHLOROTANNINS AND HYPHENATED TECHNIQUES FOR QUANTITATIVE ANALYSIS
The use of liquid chromatography (LC) to quantify the phlorotannin compositions in macroalgal extracts is limited by the lack of commercially available standards. The only standard for calibration that is commercially available is the monomer phloroglucinol.

Koivikko et al. measured the concentration of phlorotannins through integration of peaks in a crude extract of Fucus vesiculosus. 38
The F-C assay was used to calculate the TPC of these compounds and then a Pearson correlation coefficient was calculated between the individual traces of the chromatogram and the contents of total phlorotannins. Further statistical analysis was performed to assess how well the variation in the phlorotannin chromatography profile can explain the variation of the content of total phlorotannins by conducting a multiple regression analysis. This attempt at quantification is however not complete as the area of the peak in the chromatographic profile is affected by the sensitivity of the compound to detection and the stability of the compounds under analytical conditions. 38 Therefore further research needs to be performed with  56 Due to the instability of the phlorotannins, and to make the NMR analysis easier, the compounds are usually acetylated using acetic anhydride and pyridine. 25,[48][49][50][51]57 The acetylation of the hydroxyls of the phenolics prevents the keto-enol tautomerisation and thereby suppresses oxidation. In these types of phenolic compounds, the hydrogens between two phenolic groups on an aromatic ring are susceptible to hydrogen-deuterium exchange with solvents such as deuterium oxide and deuterated methanol, which can result in diminished or fully exchanged peaks in the spectra. However, acetylation of the phlorotannins prevents this exchange from occurring, also because the keto-enol tautomerisation is prevented. Before compounds can be identified by NMR spectroscopy they must first be purified by chromatography. Acetylation of the phlorotannins changes the polarity of the compounds and hence, once acetylated, the phorotannins can then be separated by normal phase silica chromatography using solvent systems of n-hexane, chloroform, methanol or ethanol. 57 Thin layer chromatography, can be used to measure the retention factor (Rf) values of the compounds with a vanillin and sulphuric acid stain for visualisation. 58 Preparative thin layer chromatography can also be used to separate compounds of small samples.    Table 6. The different monomers, from which the larger phlorotannin structures are constructed, are displayed in Figure 4 and their carbons numbered; this is not in accordance to IUPAC numbering but a system that has been used previously by Glombitza ad coworkers. 60 From Table 6 it is observed that the ether linkages correspond to peaks in the region 147.9-154.7 ppm for phenol linkages meta to an acetyl group and 136-134 ppm for phenol linkages ortho to an acetyl group. The carbon NMR data can also be seen for the Eckol structure  shown in Figure 5 the values of which are depicted in Table 7. Here it is shown that the eckol type linkages are between 130-145 ppm on a 13 C NMR spectra. From Table 7 it is observed that the eckol linkages are very close in chemical shift to the acetylated phenolics, hence 2D NMR is needed to distinguish them.  Tables 7 and 4 shows the importance of using both 1 H-and 13 C-NMR spectra to identify phlorotannin structures, either purified or in crude extracts. Using the intensity of the signals it is possible to identify the ratio of linkages and chemical shifts corresponding to certain chemical groups which can be used to determine the abundance of the different types of phlorotannin in a mixture. However, achieving good quality 13 C-NMR spectral data of a complex mixture can be difficult.

| MASS SPECTROMETRY IN QUALITATIVE ANALYSIS AND IDENTIFICATION OF PHLOROTANNINS -QUALITATIVE ANALYSIS
Mass spectrometry (MS) is a powerful tool in the analysis and identification of phlorotannin compounds. It is mostly regarded as a qualitative tool due to the lack of standards available to counteract quantitative hurdles in MS such as matrix effects. Tandem mass spectrometry or MS/MS has been frequently used coupled to various LC systems to observe the fragmentation pattern of the compounds as they elute    Normal phase chromatography involves the use of a polar stationary phase but use of this chromatography method can lead to the polar phlorotannins being difficult to elute at all due to high interaction with the stationary phase.
One study has shown separation with a hydrophilic interaction liquid chromatography (HILIC) column. 23 23 In this study the chromatograph was coupled to a high-resolution mass spectrometer in order to generate more structural information on the peaks.
More recently Ultra-performance liquid chromatography mass spectrometry (UPLC-MS) has been employed to determine the metabolite profile of the phlorotannins in three species of brown macroalgae. UPLC has been developed to endure higher system back pressures than conventional HPLC and thus much smaller column particle sizes can be used, which enhances speed and sensitivity of analysis. 66  Although HPLC cannot be used to fully identify the structure of the phlorotannin and the position of linkages, valuable information on the distribution, type and size of phlorotannins can be obtained. 8 The fragmentation behaviour of the phlorotannins can give information on the linkage type based on the charged fragments observed in the mass spectrum. 2,8 Some characteristic fragments are displayed in Figure 6. Another study separated the small (< 2 kDa) phlorotannins from Ascophyllum nodosum on a C18 column into two distinct peaks using diode array detector (DAD). 67 However, separation and identification of phlorotannins by HPLC fitted with a DAD is complicated even further by the lack of standards available for comparison. 68 Whilst DAD detection can be useful for analysis of compounds with similar molecular weights but different electronic distribution in the chromophore, leading to different UV spectra, standards or literature data are needed for comparison if the compounds are to be identified.
The composition of the mixtures in the peaks in this study were not conclusively identified due to lack of standards available. To add to this, the phlorotannin structures are very prone to oxidation during extraction due to the increased tendency of multiple phenolic groups on one aromatic ring to tautomerise to the more reactive keto form.
The instability of the phlorotannins prior to analysis can affect the reproducibility of the results when performing these experiments.
Two-dimensional LC techniques have been employed to assess the complex matrix of phlorotannin compounds in seaweed extracts from the brown seaweed Cystoseria abies-marina. 63  Colorimetric assays could be a high throughput, easy and costeffective tool for measuring phlorotannin content but robust work needs to be done in order to standardise a method to make results comparable between studies. Many papers have focused on optimising methods for one species of seaweed, but more species need to be used in these types of studies to ensure that optimised methods are suitable for most brown seaweed species.
HPLC, especially when coupled to MS, is a useful technique to study high and low molecular weight phlorotannins and their distributions in an extract. However, often specialised equipment is required to run techniques such as MALDI-ToF and HRMS which are required to detect and identify larger phlorotannins. Figure 7 is a schematic showing the sample preparation step required for the analytical techniques covered in this review. From this schematic it can be seen that the assay analysis is by far the fastest and simplest method. However,