*Article* **Optimization of an Innovative Hydrothermal Processing on Prebiotic Properties of** *Eucheuma denticulatum***, a Tropical Red Seaweed**

**Birdie Scott Padam 1, Chee Kiong Siew <sup>1</sup> and Fook Yee Chye 1,2,\***


**Featured Application: This study provides key factors for the efficient subcritical low-acid hydrolysis of oligosaccharides from red seaweed having potential prebiotic properties, a soughtafter functional food ingredient. Extraction of marine oligosaccharides from sustainable seaweed biomass with minimal loss is a green effort to cater to the huge global functional food demand.**

**Abstract:** Seaweed is a sustainable source of marine oligosaccharides that potentially could be used as a prebiotic ingredient for functional food development. The study aims to optimize the oligosaccharide preparation through thermal hydrolysis of an under-utilized red seaweed, *Eucheuma denticulatum*. Response surface methodology (RSM) applying Box–Behnken design (BBD) was used on three parameters including temperature (105–135 ◦C), hydrolysis time (15–35 min) and sulfuric acid concentration (0.05–0.2 M). Optimized fractions with good prebiotic activity were characterized using high-performance size-exclusion chromatography (HP-SEC) and Fourier transform infrared spectroscopy (FT-IR). *Eucheuma denticulatum* oligosaccharides fraction 1 (ED-F1) was shown to promote the growth of beneficial gut microbiota including *Lactobacillus plantarum*, *L. casei, L. acidophilus*, *Bifidobacterium animalis* and *B. longum* with the highest prebiotic activity score of 1.64 ± 0.17. The optimization studies showed that hydrolysis time was the most significant parameter for the oligosaccharides yield. Optimal hydrolysis conditions for ED-F1 were 120 ◦C, 21 min, 0.12 M H2SO4 with the highest yield achieved (11.15 g/100 g of dry weight). The molecular weight of ED-F1 was determined at 1025 Da while FT-IR analysis revealed the presence of sulfated oligosaccharides with similar characteristics of *i*-carrageenan. These findings signify the innovative method for the efficient production of seaweed derived prebiotic oligosaccharides, which could be a promising source of functional food ingredients for the development of health foods and beverages.

**Keywords:** optimization; response surface methodology; lactobacillus; bifidobacterium; *Eucheuma denticulatum*; seaweed oligosaccharide; HMF

## **1. Introduction**

Seaweeds are one of the most consumed marine products, especially in Asia, where it has been cultivated and used since time immemorial. It is estimated that the global seaweed industry produced 32.7 million tonnes in volume in 2020, with revenue of about USD 13.3 billion [1,2]. About 85% of the seaweed produced is directly consumed as food. These unique marine macroalgae attract consumers due to the numerous known nutritional and health benefits containing bioactive ingredients showing properties such as anti-cancer, anti-cholesterol and anti-hypertension [3–5] as well as anti-inflammatory [6]. It is also reported to be a sustainable source for potential prebiotics [7,8]. Among the many components available from seaweeds, the polysaccharides and oligosaccharides from these

**Citation:** Padam, B.S.; Siew, C.K.; Chye, F.Y. Optimization of an Innovative Hydrothermal Processing on Prebiotic Properties of *Eucheuma denticulatum*, a Tropical Red Seaweed. *Appl. Sci.* **2023**, *13*, 1517. https:// doi.org/10.3390/app13031517

Academic Editor: Monica Gallo

Received: 20 December 2022 Revised: 17 January 2023 Accepted: 19 January 2023 Published: 24 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

marine biomasses are known to be unique and have gained a tremendous research interest for food and pharmaceutical applications.

The diverse number of species within the three major groups of seaweed (red, green and brown), provides a huge variation in the types and properties of their polysaccharides (carrageenan, alginate, agar) which have different applications in the food industry [9]. Polysaccharides are generally deposited in the seaweed cell matrix, mainly for structural purposes. Conventionally, polysaccharides are extracted from seaweeds through physical and chemical degradation of biomass tissues using the hydrothermal method, alkaline and acid hydrolysis, as well as enzymatic preparation, while further degradation of the extracted polysaccharide or prolonged degradation of the seaweed biomass yields low molecular weight polysaccharides and oligosaccharides [10]. Some seaweed polysaccharides are heavily sulfated and vary tremendously in their monosaccharide composition as well as polymer linkage [11]. A typical polysaccharide such as agarose has an average molecular weight of more than 100,000 Da and 0.15% sulfate content, while a low molecular weight agaropectin has a mass below 20,000 Da with up to 8% sulfate content [12].

Current studies showed that some seaweeds' non-digestible polysaccharides and hydrolysed low molecular weight polysaccharides and oligosaccharides promote the growth of beneficial gut microbiota such as lactobacilli and bifidobacterium. These include alginate and agar polysaccharides from *Grateloupia filicina*, *Eucheuma spinosum, Kappaphycus alvarezii* [13,14] as well as commercially available laminaran (*Laminaria* sp.), ulvan (*Ulva* sp.) and porphyran (*Propia* sp.) [11]. Agaro-oligosaccharides derived from the hydrolysis *Gracillia lemaneiformis* polysaccharides also exhibit similar activity in promoting significant abundance in beneficial microbiota population while *Sargassum confusum* oligosaccharides display potential anti-diabetic properties through regulating the gut microbiota [15,16]. Low molecular weight oligosaccharides with such activities are much sought after not only for their physicochemical properties, but also for their bioactivity, with the potential to create huge impact in the industry [17].

Extraction of oligosaccharide is the first and most challenging step in determining the composition and properties of the compounds, in which the recovery is affected by several factors such as temperature, solvent, and extraction technique. Conventional extraction methods using high temperatures for a prolonged period is always detrimental to sample as the decomposition of polysaccharide backbone yields by-products such as hydroxymethyl furfural and furfural derivatives that can be toxic at high concentrations [18,19]. Diluted acid hydrolysis coupled with high temperature extraction by far is still the low-cost method for the recovery of oligosaccharides from lignocellulosic material. However, there is very little study of the utilization of low acid hydrolysis and its effects on the degree of recovery of oligosaccharides from seaweeds, especially for the less-cultivated varieties.

Response surface methodology (RSM) through Box–Behnken design (BBD) is a powerful tool to predict the best extraction conditions leading to optimal desired experimental responses. Recent published literatures have demonstrated this method combination for the efficient extraction of seaweed phytosterols [20], plant antioxidants [21] and mushroom peptides [22]. A minimum of three levels of each shortlisted parameter is required for the BBD to fit a second-order regression model (quadratic model) [21]. In this context, the current study aimed to investigate the effects of temperature, time and acid concentration on the oligosaccharides extraction from red seaweed *Eucheuma denticulatum* and optimizing the extraction parameters using RSM for higher recovery of the oligosaccharide fractions while reducing the formation of by-product 5-hydroxymethyl furfural (HMF).

#### **2. Materials and Methods**

#### *2.1. Materials*

All solvents, reagents, chemical standards and microbiological medium were purchased from Merck (Darmstadt, Germany) unless otherwise stated. *Lactobacillus paracasei* (LC01), bifidobacterium lactis BB12 (BB12) and *Bifidobacterium longum* BB46 (BB46) cultures were from Chrs. Hansen (Hørsholm, Denmark). *Lactobacillus plantarum* ATCC 8014

(LP8014), and *Escherichia coli* ATCC 11775 (EC11775) were from American Type Culture Collection (ATCC, Manassas, VA, USA). Commercial prebiotic (Fibrulose F97™) was from Cosucra (Pecq, Belgium).

## *2.2. Seaweed Sample Preparation*

Red seaweed (*Eucheuma denticulatum*) was collected from Semporna, Sabah Malaysia with the assistance from the Seaweed Research Unit, Universiti Malaysia Sabah. Seaweeds were washed from impurities, dried at 50 ◦C until reaching approximately 10% moisture content, milled and were kept in an airtight container in −20 ◦C. Species verification was performed by the Seaweed Research Unit, Universiti Malaysia Sabah.

## *2.3. Hydrolysis and Extraction of Seaweed Oligosaccharide Fraction*

The seaweed oligosaccharide fraction was extracted using the modified autoclave method (130 ◦C, 15 psi, 15 min) (Hirayama, Japan) [23] with low acid hydrolysis (0.2 M H2SO4). Dried and defatted seaweed powder were used (25 g, 1:20 solid–liquid ratio) and placed in 1 L blue-cap Schott™ bottles. Recovered liquor was neutralized with calcium carbonate and the supernatant was recovered through centrifugation (5000 rpm, 5 min). The supernatant was concentrated under vacuum until 10% of its initial volume at 60 ◦C and was deproteinized through the removal of precipitates (8000 rpm, 5 min). Sequentially, two volumes and five volumes of ethanol (95%) were added to the supernatant under constant stirring and both precipitates were removed through centrifugation. The supernatant obtained after the removal of the ethanol precipitate were concentrated again to 5% volume and were precipitated using nine volumes of pure ethanol (99.9%) to obtain low molecular weight oligosaccharides. Oligosaccharide precipitates were washed with pure ethanol twice and vacuum dried at 50 ◦C. Bradford's protein assay was used to check for residual protein. Oligosaccharide precipitates were fractionated through anion exchange column (DEAE-Sepharose® Fast Flow Column, 20 mm × 100 cm) (Sigma-Aldrich, St. Louis, MO, USA) and gradient eluted with distilled water, 0.1, 0.2, 0.4 M, 0.8 M and 1.6 M NaCl solutions [24]. Each fraction was collected and the phenol–sulfuric acid assay [25] was used to determine the sulfated sugar content. Fractions which contained oligosaccharides were pooled, concentrated, desalted (Sephadex® G-10 column, 2.0 × 25 cm) (Sigma-Aldrich, St. Louis, MO, USA), vacuum dried and weighed. Fractions were subjected to size exclusion chromatography (HP-SEC) (Agilent 1200, Santa Clara, CA, USA) using BIOSEP-SEC S2000 column (Phenomenex, Torrance, CA, USA). Elution took place at 30 ◦C with 50 mM sodium nitrate and the elution was monitored using a refractive index detector (RID) [26]. Calibration was performed using galactose, maltose, maltotriose, and dextrans with molecular weight ranging between 1 and 25 kDa as standards.

#### *2.4. Prebiotic Activity Assay and Prebiotic Activity Score*

The assay was conducted by adding bacterial suspensions (final inoculum concentrations of 6 log CFU/mL) to separate tubes containing MRS basal broth (probiotic) or M9 broth (EC) (with 1% (wt/vol) glucose or 1% (wt/vol) prebiotic. Cultures were incubated under anaerobic conditions at 37 ◦C. After 0 h and 24 h of incubation, bacterial samples were enumerated on MRS and TSA, respectively. Each assay was performed in triplicate. The prebiotic activity score was calculated using the Equation (1) according to Huebner et al. 2007 [27]:

Prebiotic activity score = probiotic (Δ log CFU/mL prebiotic/Δ log CFU/mL glucose) − enteric (<sup>Δ</sup> log CFU/mL prebiotic/<sup>Δ</sup> log CFU/mL glucose) (1)

## *2.5. Quantification of HMF*

Briefly, the chromatographic separation was performed using an Agilent 1200 Series HPLC, equipped with a diode array detector (DAD) (Agilent, Santa Clara, CA, USA). Extraction liquor containing 5-hydroxy-2-methylfurfural (HMF) was injected through a

Phenomenex column (Luna C18, 5.0 μm, 4.6 × 150 mm) at 30 ◦C. The mobile phase consisted of water (with 0.5% formic acid) and acetonitrile in the ratio 90:10 (vol/vol) under isocratic conditions, at a flow rate of 0.8 mL/min with an injection volume of 30.0 μL. The detection was carried out at 285 nm and the total run time was 8 min. All aqueous samples solutions were filtered on 0.45 μm nylon filters before injection on chromatographic system [28].

## *2.6. Selection of Three Levels of Independent Variables (X1, X2, X3)*

Extraction of oligosaccharide fraction ED-F1 was done at different temperatures (95 ◦C to 135 ◦C hours) at constant time (15 min) and 0.2 M H2SO4 concentration. The best extraction temperatures were forwarded down and were used as a constant for the next extraction procedure in which the time is varied (15 min to 55 min) while applying a constant 0.2 M H2SO4 concentration. The best extraction temperatures applied with the best extraction time were forwarded down as constant variables for the next extraction procedure varying the H2SO4 concentration (0.05 M to 0.25 M). All extraction liquors were similarly purified according to the previous method to obtain fraction ED-F1. Parameters with the highest fraction yield were selected as center points.

## *2.7. Experimental Design and Validation of Models*

The optimization approach was carried out using The Design Expert (Version 11.0.0, Stat-Ease Inc, Minneapolis, MN, USA) according to a three level, three variable Box– Behnken design with 17 design points. Three independent variables consist of hydrolysis temperature (◦C, X1), time (min, X2) and sulfuric acid concentration (M H2SO4, X3). A narrowed three levels of each variable were selected based on the results from the single factor experimentation, denoted as lower level (−1), upper level (+1) and including the center point (0). Levels were assigned accordingly for X1: 115 ◦C(−1), 125 ◦C(0), to 135 ◦C(+1); X2: 15 min(−1), 25 min(0), 35 min(+1); X3: 0.05 M(−1), 0.125 M(0), 0.2 M(+1). Responses (Y) are based on the recovered yield of oligosaccharide fraction in grams and hydroxymethyl furfural concentration in the reaction liquor (g/L). Experimental data were fitted to the following second order polynomial equation proposed for the analysis of each response (Y) shown by Equation (2):

$$\text{Y = \beta 0 + \sum\_{i=1}^{3} \beta \text{i}\lambda \text{i} + \sum\_{i=1}^{3} \beta \text{ii}\lambda \text{i}^2 + \sum\_{i \neq j=1}^{3} \beta \text{ij}\lambda \text{i}\lambda \text{j} \tag{2}$$

In the equation, Y is the yield of oligosaccharide fraction and hydroxymethyl furfural concentration, predicted response; β0, βi, βii, and βij are regression coefficients for intercept, linear, quadratic and interaction terms respectively; Xi and Xj are independent variables. The significance of the model was calculated in terms contributing to the regression sum of squares. The reduced model was then acquired through the exclusion of the non-significant coefficients from the initial model after the analysis the regression model coefficients (R2) and evaluating the model lack of fit using ANOVA (*p* < 0.05) and the Fisher test value (F-value). Response surface plots were developed to explain the effects of independent variables (temperature, time and H2SO4 concentration) on the response variables (ED-F1 fraction, HMF) [29].

## *2.8. FT-IR Spectra Acquisition*

The Fourier transform infrared (FT-IR) spectrum of the ED-F1 was detected on the FT-IR spectrometer (Spectrum 100, Perkin Elmer, Waltham, MA, USA), recorded in a transmittance mode over a wavelength range between 4000 and 400 cm−<sup>1</sup> [30]. Iotacarrageenan was used as a reference standard. Triplicates of each sample were scanned to get an average spectrum.

#### *2.9. Statistical Analysis*

All determinations were performed at least in triplicate. Data were expressed as mean values ± standard deviation. Comparison of means was performed by one-way analysis of variance (ANOVA) with a significance level of *p* < 0.05. The Design Expert software (ver.12) (Stat-Ease, Inc., Minneapolis, MN, USA) was used for constructing the regression model, designing the Box–Behnken and predicting the optimal parameters.

## **3. Results**

#### *3.1. Purification and Prebiotic Activity Score of Red Seaweed Oligosaccharide Fractions*

Red seaweed oligosaccharide fractions undergoing anion-exchange column yielded three distinct fractions assigned ED-F1, ED-F2 and ED-F3 (Figure 1). The first fraction eluted out with water resulted in the major peak, followed by a peak eluted using 0.2 M and 0.4 M NaCl, while no peak was detected through the elution using 0.8 M–1.6 M (results not shown). Monitoring was conducted using the phenol–sulfuric assay and similar fractions were pooled together into one single fraction. Anion-exchange chromatography was extensively used to fractionate oligosaccharides including fucooligosaccharides derived from the enzymatic hydrolysis of brown seaweed Sargassum honeri [31] as well as ulvan oligosaccharides from green seaweed *Ulva* sp. [32]. Gradient increase in the ionic strength of the mobile phase (Cl−), competes with the binding site of the positively charged DEAE sepharose resin that will gradually enable the release of negatively charged molecules that binded earlier [33]. Ionized sulphate ester groups are the prime contributors to the natural anionic properties of the seaweed oligosaccharides [34].

**Figure 1.** *E. denticulatum* low molecular weight oligosaccharide fractions obtained through anion exchange chromatography (DEAE-Sepharose® Fast Flow).

The three oligosaccharide fractions subjected to prebiotic assay showed positive prebiotic activity score values against five different commercial probiotics (Figure 2). ED-F1 displayed the highest prebiotic activity score against all probiotics with the highest value against *L. paracasei* LC01 at 1.64 ± 0.17. The prebiotic activity score values of ED-F1 were significantly higher compared to commercial prebiotic F97 against *B. animalis* BB12, *B. longum* BB46, *L. paracasei* LC01 and *L. acidophilus* LA05. These values are higher than the reported prebiotic activity score of pectic oligosaccharide fractions from citrus peel against *L. paracasei* (0.17–0.38) and *Bifidobacterium bifidum* (0.09–0.93) [35], but comparable to the score of seaweed polysaccharides extracted from *Sargassum withii* (1.42) and *Enteromorpha compressa* (1.44) against *Lactobacillus plantarum* [36]. The current study displayed a slightly lower prebiotic activity score values of ED-F1 against *L. plantarum* LP8014. The variation observed in the prebiotic activity score of different fractions for both bifidobacteria and lactobacilli strain relates to its metabolic diversity [27] and their different preferential utilization of various oligosaccharides in the fraction [27,37,38]. Depending on the species, lactobacilli strains have been known to possess a great variety of genes involved in the metabolism of both complex oligosaccharides and simple sugars which can be switched on depending on the availability of the type of carbohydrates available [39].

**Figure 2.** Prebiotic activity score of *E. denticulatum* fractions obtained through anion exchange chromatography (ED-F1, ED-F2, ED-F3) compared against commercial prebiotic Fibrulose (F97) and a negative control (Cellulose).

## *3.2. Selection of Factor Levels*

Both ED-F1 fraction yield and HMF concentration in the extraction liquor showed a positive linear relationship affected by extraction temperature from 95 ◦C to 135 ◦C (Figure 3a) with the yield of ED-F1 fraction increased by 6.52 g/100 g at extraction temperature of 125 ◦C compared to 95 ◦C. It has been shown previously that water at subcritical level (>100 ◦C) has the tendency to produce hydronium ions (H3O+), which resulted in the rapid hydrolysis of macromolecules into smaller molecules [40]. It was observed that the increment of temperature in the hydrolysis of passion fruit peel and oat is proportionate to the gradual increase in solid loss percentage of the raw material as well as the increase in detectable smaller carbohydrates such as oligosaccharides and monosaccharides as well as acids and HMF by-products [21,40]. For optimization, a shorter temperature range from 115 ◦C to 135 ◦C was selected based on significance increase in the yield of ED-F1 fraction at temperatures 125 ◦C to 135 ◦C. The temperatures 115 ◦C and 135 ◦C were selected for the lower and upper level, respectively, to be applied in the optimization design using RSM.

**Figure 3.** The yield of ED-F1 fraction and HMF by-product as affected by (**a**) hydrolysis temperature, (**b**) hydrolysis time and (**c**) H2SO4 concentration.

When the temperature of 125 ◦C was brought down for further application in varying the time of hydrolysis, the results were observed showing an inverse linear relationship between the yield of ED-F1 fraction and hydrolysis time 25 to 55 min while HMF increased gradually from 15 to 55 min. Liu et al. (2020) [41], showed that the degree of hydrolysis of fucosylated glycosaminoglycan using mild acid increase exponentially against time at temperatures below 100 ◦C over the span from 1 to 36 h while Sophonputtanaphoca et al. (2018) [42], reported that oligosaccharides can be produced within 30 to 60 min at temperatures between 100–121 ◦C. In the current study, a gradual decrease of ED-F1 yield observed at temperature 125 ◦C against a timespan of 35 to 55 min suggested a preferable shorter hydrolysis time. The gradual decrease of ED-F1 could be due to the hydrolysis of the oligosaccharide components generating more HMF as shown in (Figure 3b). It is also important to note that time of hydrolysis may also be affected by the plant matrix itself made up of different glycan backbones between the current and previous studies. Time of hydrolysis at 15 min to 35 min was selected based on the observed peak increase of the ED-F1 fraction yield at 25 min.

Incrementing the H2SO4 concentration in the hydrolysis solvent showed an increasing trend of ED-F1 fraction yield with a plateauing trend at 0.2 M–0.25 M concentration at constant 25 min time (Figure 3c). In contrast, Wang et al. (2019) [43], demonstrated that mild sulfuric acid concentration (0.3–0.7 M) had a significantly increasing trend in the hydrolysis of press-lye waste hemicelluloses to xylo-oligosaccharides. This could be explained by the difference in the other factors used such as hydrolysis time and temperature as well as the difference in the type of carbohydrates in the raw material in the current study. It is also shown in the previous study by Sophonputtanaphoca et al. (2018), acid concentration beyond a certain range with the effect of time and temperature favours the production of monosaccharide over the oligosaccharide fraction with the oligosaccharide completely undetectable at hydrolysis using sulfuric acid concentration at 2 M for 1 h [42]. A range of 0.05 M to 0.2 M H2SO4 concentration was selected for the lower and upper level for optimization.

#### *3.3. RSM Model and Analysis of Variance*

RSM was used to investigate the effects of hydrolysis temperature (115–135 ◦C), extraction time (15 min–35 min) and H2SO4 concentration (0.05 M–0.2 M) on the oligosaccharides fraction ED-F1 with potential prebiotic from *E. denticulatum*. Several studies have successfully employed this method in optimizing hydrolysis parameters to obtain corncob xylooligosaccharides from a microwave-assisted hydrolysis [44], deriving glucomannooligosaccharides from copra meal hydrolysis [45] and enzymatic hydrolysis to obtain galacto (arabino)-oligosaccharide from potato rhamnogalacturonan [46]. ED-F1 fraction was selected to be optimized due to its significant and highest prebiotic activity score against five commercial probiotics. HMF was also measured as it is directly associated with the decomposition of hexoses under heat treatment and was known to be a potential carcinogen at high amounts.

Three hydrolysis parameters were selected as factors and the operating range were narrowed. These parameters have been shown to significantly affect ED-F1 fraction yield. The factors were assigned as hydrolysis temperature (X1), hydrolysis time (X2) and H2SO4 concentration (X3) (Table 1). Several researchers also have emphasized these to be vital factors that highly affect low or mild acid hydrolysis procedures to obtain oligosaccharides [44,47,48]. The Box–Behnken model was selected for the design of the experiment considering the restriction of operability in the selected factor (maximum heating temperature of the autoclave). Not only does the Box–Behnken design accommodate for a minimum number of experiments, but it also efficiently allows estimating first order and second order coefficients of the model and provides supportive analysis on the interactions between the variables [49]. The three selected factors were assigned a lower, center point and upper level designated in code of −1, 0, +1, respectively, in which the central point and the experimental points are equidistant.

From the 17 experiments, a quadratic model was adjusted to the responses (ED-F1 and HMF), and the regression coefficients for the linear, quadratic, and interaction terms were calculated and statistically evaluated using ANOVA. Table 2 shows the statistical analysis of the regression coefficients of the complete polynomial models for the yield of ED1-F1 and HMF. Based on the ANOVA (*p* ≤ 0.05), the linear first-order effect was significant for all factors (temperature X1, time X2, H2SO4 concentration X3), the quadratic second-order effect was significant only for temperature and H2SO4 concentration for both ED-F1 yield and HMF.


**Table 1.** Box–Behnken experimental design layout and observed responses of oligosaccharide fraction yield (ED-F1) and HMF by-product.

<sup>1</sup> <sup>X</sup> <sup>1</sup> Temperature: hydrolysis temperature; X2 Time: hydrolysis time and X3 H2SO4: sulfuric acid concentration. <sup>2</sup> ED-F1: *E. denticulatum* fraction 1; HMF: 5-hydroxymethyl furfural.

**Table 2.** Regression coefficients for the polynomial model for oligosaccharides hydrolysis from *E. denticulatum*.


<sup>1</sup> ED-F1: *E. denticulatum* fraction 1; HMF: 5-hydroxymethyl furfural. \*\*\* *p* < 0.001, \*\* *p* < 0.01, \* *p* < 0.05.

The effect of the interaction between any two factors (X 1, X2 and X 3) was significant for ED-F1 yield. Similar occurrence was observed for HMF by-product except for the non-significance between interactions of X1X3. The insignificant coefficients were removed from the second-order polynomial model.

Thus, the reduced regression equation with the coded values was established as the following:

$$\text{ED-F1 yield (g)} = 9.97 - 0.098\text{X}\_1 - 0.216\text{X}\_2 - 0.121\text{X}\_3 - 0.722\text{X}\_1^2 - 1.19\text{X}\_3^2 - 0.365\text{X}\_1\text{X}\_2 - 0.34\text{X}\_1\text{X}\_3 - 0.158\text{X}\_2\text{X}\_3; \tag{3}$$

HMF by-product (g/L) = 5.88 + 0.72X <sup>1</sup> + 0.39X <sup>2</sup> + 1.37X3 + 0. 90X1 <sup>2</sup> <sup>−</sup> 1.05X3 <sup>2</sup> + 0.23X 1X <sup>2</sup> <sup>−</sup> 0.0063X2X3; (4)

where X1 is hydrolysis temperature, X2 is hydrolysis time and X3 is H2SO4 concentration.

The reduced quadratic models were statistically significant with the *p*-value for both responses (ED-F1 yield and HMF) shows <0.0001, indicating the model is statistically significant at 95.0% confidence interval (*p* < 0.05). Furthermore, the aptness of the secondorder polynomial model was confirmed by the insignificant lack of fit F-values (*p* > 0.05), indicating the lack of fit is not significant relative to the pure error in this experiment [50]. The regression coefficient (*R*2) was well adjusted for the experimental data (ED-F1 yield adjusted *R*<sup>2</sup> = 0.9836, HMF adjusted *R*<sup>2</sup> = 0.9616) which indicates that the quality of the models was retained even after the removal of some terms. Based on the adjusted *R*<sup>2</sup> value for both models, only 1.74% and 3.94% of the total variation is not explained by both ED-F1 yield and HMF models, respectively, with the variations in the ED-F1 yield (98.36%) and HMF content (96.16%) are direct results to changes in the temperature, time and H2SO4 concentration. The difference between the predicted *R*<sup>2</sup> and the adjusted *R*<sup>2</sup> for both models were in reasonable agreement (less than 0.2) while the model fitting was better as R<sup>2</sup> value observed is closest to 1 [51]. Moreover, the lower value of CV (<5) for ED-F1 implies the low deviations between the experimental data values and the predicted data values. Models with low CV values (<10) are still considered to have good reliability and reproducibility [52]. The "Adeq Precision" which implies the signal-to-noise ratio, indicates an adequate signal (>4) for all models and can be used to navigate the design space [49,50,53].

## *3.4. Effects of Hydrolysis Parameters on the Yield of Oligosaccharides Fraction ED-F1 and HMF By-Products*

Based on Table 1, ED-F1 yield ranged from 7.65 to 10.11 g according to the changes in the levels of extraction parameters. The lowest yield was obtained under the conditions of X1: 135 ◦C, X2: 25 min and X3: 0.2 M H2SO4, while the highest yield was observed in X1: 125 ◦C, X2: 25 min and X3: 0.13 M H2SO4. Time of extraction (X2) (β: −0.216) exhibited the highest significance (*p* < 0.001) for the yield of ED-F1 as compared to both temperature (X1) (β: −0.098) and H2SO4 concentration (X3) (β: −0.121) (*p* < 0.05, respectively), while for the HMF production in the extraction liquor, all three parameters are equally significant (*p* < 0.001) with all positive values of the regression coefficients (Table 2). Positive regression coefficients imply that the increment of the parameter levels will have a direct proportion to the production of the product, vice-versa [54]. Interaction between any two parameters X1, X2 and X3 are highly significant to the yield of ED-F1 (*p* < 0.001), but on the opposite, interaction of X1-X3 is not significant for the formation of HMF. Similar occurrence was reported by [55] in optimizing the hydrolysis conditions for xylans from beech wood and corn cob. Their results indicated that temperature and time of hydrolysis plays a pivotal factor in the conversion of hemicelluloses to oligosaccharides using the minimum concentration of sulfuric acid, thus reducing the degradation of xylose monomers to furfural by-product.

Temperature and H2SO4 concentration also showed a significant effect on the hydrolysis yield of ED-F1 mainly with the temperature around 120–130 ◦C and H2SO4 concentration around 0.9–1.6 M (Figure 4). Temperatures below 120 ◦C and higher than 130 ◦C showed reduction in hydrolysis yield of ED-F1 which can be confirmed by the negative quadratic

term of temperature in the reduced mathematical model. Similarly, H2SO4 concentration range below 0.09 M and higher than 0.16 M showed negative impact on ED-F1 yield. Subcritical temperatures beyond the boiling point of water with the addition of mild acid have been shown to efficiently reduce the hydrolysis time for a higher yield of oligosaccharide fraction. According to Wang et al. (2018), the highest yield (45.18%) of xylooligosaccharide (XOS) fraction using subcritical water with 1% (0.1 M) sulfuric acid from hemicellulose was the highest at 120 ◦C for 60 min [56]. Temperatures beyond 120 ◦C have been noted to produce more xylose monomers. This observation could be explained by the different hydrolysis rate constant of the substrate at different temperature and acid concentration that affects the scission of terminal-nonreducing bonds, interior glycosidic bonds and terminal-reducing C-O bonds; the chain length of the produced oligosaccharides and the constant have been shown to be proportionally related [57].

**Figure 4.** Three-dimensional response surface plots showing the effects of temperature, time and H2SO4 on ED-F1 yield (**a**–**c**) and HMF (**d**,**e**).

## *3.5. Validation of the Model and Optimal Extraction Condition of ED-F1 Oligosaccharide Fraction*

Based on the response surfaces, the yield of ED-F1 is highly influenced by the hydrolysis parameters. Hence, it is important to get the most desirable parameter conditions to obtain the optimum ED-F1 yield. Since HMF production is directly proportional to the reduction of ED-F1, the goal is set to reduce the production of HMF in the extraction liquor. The reduced quadratic models were used to generate optimal hydrolysis conditions for ED-F1 oligosaccharide fraction yield and the values were 121 ◦C, 21 min and 0.12 M H2SO4 concentration (Table 3). The desirability value was close to 1 and calculated to be 0.844. All the hydrolysis parameters were set in range while the ED-F1 fraction yield and HMF concentration were both set at maximum and at minimum, respectively. The experimental validation of the mathematical model was performed in triplicate. The ED-F1 yield obtained through the optimization procedure was 11.15 ± 0.03 g/100 g dw. Both ED-F1 yield and HMF by-product produced values were close to the predicted by the mathematical model in optimal conditions with a low percentage of relative error (3.15% and 4.72%, respectively) reflects the adequacy of the developed quadratic models [52]. The optimization procedure increased the yield of ED-F1 by 31.41% compared to initial extraction and hydrolysis methods.


**Table 3.** Validation of conditions for the optimum yield of ED-F1.

<sup>1</sup> X1 Temperature: hydrolysis temperature; X2 Time: hydrolysis time and X3 H2SO4: sulfuric acid concentration. <sup>2</sup> ED-F1: *E. denticulatum* fraction 1; HMF: 5-hydroxymethyl furfural.

## *3.6. Characterization of ED-F1*

3.6.1. Size-Exclusion Chromatography

ED-F1 was subjected to HP-SEC to determine the homogeneity of the fraction as well as its molecular weight. Figure 5 shows a single symmetrical peak eluted at 13.84 min depicting that the sample is highly homogenous containing oligosaccharides of similar or close in terms of their molecular weight [58]. The regression equation obtained from the oligosaccharide and dextran standards was LogMW = −0.2252X + 6.1281 with a correlation coefficient (R2) of 0.9826. The average molecular weight was represented by Mw, and the elution time was represented by X. According to the equation, the average molecular weight of ED-F1 was calculated as 1.025 × <sup>10</sup><sup>3</sup> Da. Low molecular weight manno-oligosaccharides and galacto-oligosaccharide (<1.0 KDa) was shown to promote the growth of *Bifidobacteria* and *Lactobacilli* in an in vitro fermentation [59]. Based on the earlier study, these low molecular weight oligosaccharides are easily metabolized in a pure culture study and efficiently being converted to lactate by *Lactobacilli* species. Oligosaccharides derived from agaran and carrageenan with the molecular weight in the range of 0.4–1.4 and 1.0–7.0 KDa has also been shown to promote the beneficial gut microflora of pigs namely Ruminococcaceae, *Coprococcus*, *Roseburia*, and *Faecalibacterium* [60].

**Figure 5.** *E. denticulatum* ED-F1 fraction observed under size-exclusion chromatography (HP-SEC).

3.6.2. Fourier Transform Infrared Spectra Analysis

ED-F1 was subjected to FT-IR spectroscopy to determine the characteristic absorption bands and *i*-carrageenan was used as a standard. In both samples, two similar bands were observed in the 4000–2000 cm−<sup>1</sup> region of the FT-IR spectra (Figure 6). The broad absorption appeared at 3200–3500 cm−<sup>1</sup> represents hydrogen bonded O–H stretching vibrations while a weak signal at 2926 cm−<sup>1</sup> is due to C–H stretching vibrations. The FT-IR absorption band observed around the region of 1075–1041 cm−<sup>1</sup> was attributed to glycosidic linkages connecting sugar molecules and a common trait to all polysaccharides and oligosaccharides. The band around 930 cm−<sup>1</sup> were ascribed to the vibration of C–O–C bridge of 3,6-anhydrogalactose. These bands were also observed in earlier studies in both agaro-oligosaccharides from enzymatic hydrolysis of agar [61], sulfated polysaccharides extracted from red seaweeds [62]. The 3,6-anhydrogalactose polymer unit is one of the unique features that distinguished agaran and carrageenan type oligosaccharides from other common seaweed derived polymer such as alginate type oligosaccharide [63,64]. However, the 930 cm−<sup>1</sup> peak is known to be absent in both *mu* and *lambda* carrageenan as the C-O-C bridge is replaced by a sulfate group RO-SO3 [64].

**Figure 6.** FT-IR spectroscopy of *E. denticulatum* ED-F1 fraction in comparison with *i*-carrageenan standard (4000–400 cm<sup>−</sup>1).

Sulfation observed in ED-F1 was similar to the *i*-carrageenan standard designated at C4 of the galactose unit (841 cm−1) and C2 of the 3,6-anhydrogalactose (805 cm−1) [63]. A small peak near 1370 cm−<sup>1</sup> also indicates the presence of sulfate groups in both samples. A broad IR peak was also observed between 1120 and 1270 cm−<sup>1</sup> and is indicative of the S=O stretching vibration of sulfate groups on both sample [62]. A slightly higher intensity of 890 cm−<sup>1</sup> peak was observed in ED-F1 but almost unobservable at *i*-carrageenan. This peak represents the stretching vibration of the anomeric C-H of unsulfated β-galactopyranosyl residues implying the presence of this residue in ED-F1 but very minimal in *i*-carrageenan. A previous study by [60] compared the prebiotic activity of sulfated and non-sulfated seaweed derived oligosaccharides namely alginate oligosaccharide, agaro- oligosaccharide and *k*-carrageenan oligosaccharides. Sulfated and non-sulfated oligosaccharides have been shown to have distinct effects on the gut microbiota in terms of the types of beneficial bacteria groups promoted, opportunistic pathogen population and production of different ratios of short chain fatty acids (SCFAs). Thus, the prebiotic efficacy of ED-F1 having both sulfated and non-sulfated β-galactopyranosyl residues in comparison with other either sulfated or non-sulfated marine oligosaccharides need further verification.

## **4. Conclusions**

*E. denticulatum* seaweed has been shown to be a potential source of low molecular weight prebiotic oligosaccharides with the observed prebiotic activity scores of ED-F1 (against five different probiotic) that are significantly higher compared to a commercial prebiotic. Thermal hydrolysis with low acid concentration has been proven to be a suitable method to derive these prebiotic oligosaccharides from the macroalgae matrix. The temperature, time of hydrolysis and H2SO4 concentration significantly affected the hydrolysis of *E. denticulatum* cell matrix. Based on the regression coefficient and *p*-value of all models, time of hydrolysis (X2) was shown to be the most significant factor in determining the yield of ED-F1 oligosaccharide fraction while the further degradation towards the production of by-product HMF was equally affected all factors. Box–Behnken design and response

surface methodology was successfully employed to optimize the hydrolysis procedure and the maximum yield of oligosaccharides ED-F1 fraction from *E. denticulatum* was derived under the following conditions: temperature 121 ◦C, time 21 min and 0.12 M H2SO4. The optimized parameter yielded 11.15 g of ED-F1 fraction per 100 g of dry seaweed material. Initial characterization of the fraction revealed a composition similar to a carrageenan type oligosaccharide. Results obtained from this study further support the possibility of utilizing an underutilized seaweed as a potential source of a functional ingredient in the food industry with a rapid hydrolysis and extraction process potentially applicable to other macroalgae species. However, further work needs to be done in order to identify individual oligosaccharide components within the fraction as well as their prebiotic efficacy in an animal model.

**Author Contributions:** Writing—original draft preparation, B.S.P.; conceptualization, methodology, supervision, F.Y.C.; writing—review and editing, C.K.S. and F.Y.C.; project administration and funding acquisition, F.Y.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was financially supported by the Ministry of Higher Education, Malaysia, Grant Number ERGS0039-STWN-1/2013.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
