**Management of Plant Physiology with Beneficial Bacteria to Improve Leaf Bioactive Profiles and Plant Adaptation under Saline Stress in** *Olea europea* **L.**

#### **Estrella Galicia-Campos, Beatriz Ramos-Solano, Mª. Belén Montero-Palmero, F. Javier Gutierrez-Mañero and Ana García-Villaraco \***

Universidad San Pablo-CEU Universities, Facultad de Farmacia, Ctra Boadilla del Monte km 5.3, 28668 Boadilla del Monte, Madrid, Spain; e.galicia@usp.ceu.es (E.G.-C.); bramsol@ceu.es (B.R.-S.); mariabelen.monteropalmero@ceu.es (M.B.M.-P.); jgutierrez.fcex@ceu.es (F.J.G.-M.) **\*** Correspondence: anabec.fcex@ceu.es; Tel.: +34-91-3724785

Received: 8 November 2019; Accepted: 27 December 2019; Published: 7 January 2020

**Abstract:** Global climate change has increased warming with a concomitant decrease in water availability and increased soil salinity, factors that compromise agronomic production. On the other hand, new agronomic developments using irrigation systems demand increasing amounts of water to achieve an increase in yields. Therefore, new challenges appear to improve plant fitness and yield, while limiting water supply for specific crops, particularly, olive trees. Plants have developed several innate mechanisms to overcome water shortage and the use of beneficial microorganisms to ameliorate symptoms appears as a challenging alternative. Our aim is to improve plant fitness with beneficial bacterial strains capable of triggering plant metabolism that targets several mechanisms simultaneously. Our secondary aim is to improve the content of molecules with bioactive effects to valorize pruning residues. To analyze bacterial effects on olive plantlets that are grown in saline soil, photosynthesis, photosynthetic pigments, osmolytes (proline and soluble sugars), and reactive oxygen species (ROS)-scavenging enzymes (superoxide dismutase-SOD and ascorbate peroxidase-APX) and molecules (phenols, flavonols, and oleuropein) were determined. We found photosynthetic pigments, antioxidant molecules, net photosynthesis, and water use efficiency to be the most affected parameters. Most strains decreased pigments and increased osmolytes and phenols, and only one strain increased the antihypertensive molecule oleuropein. All strains increased net photosynthesis, but only three increased water use efficiency. In conclusion, among the ten strains, three improved water use efficiency and one increased values of pruning residues.

**Keywords:** olive; salinity; osmolytes; adaptation; secondary metabolism; plant growth promoting rhizobacteria (PGPR); net photosynthesis; oleuropein; water use efficiency (WUE)

#### **1. Introduction**

The traditional olive production system in the Mediterranean was developed in dry, farmed areas with trees spaced from 25 to 60 feet (7.6–18.3 m) apart, giving 12 to 70 trees/acre (30–173 trees/ha) [1]. Thus, olive trees are often under severe water deficit combined with high temperatures and high light intensities during the summer season. Moreover, the traditional olive production system in dry farms has many disadvantages such as low yields, delays before full production (15–40 years) and a very inefficient non-mechanical harvest [1]. In recent decades, the cultivation of the Spanish olive has undergone major technological changes associated with high-density or super high-density production systems, such as a reduction in the number of olive varieties, an increase in the density of the new plantations, an improvement of the harvesting machinery or orchard irrigation [2]. Therefore, this change from dry, farmed areas to irrigated cultivation has placed water stress or high salt

concentration, as one of the main problems that olive cultivation is currently facing. In this agronomic framework, in which olive trees require irrigation, the possibility of producing in areas with high salt concentration or reduced water supply represents an important economic advantage.

Although olive resists a high degree of drought stress, the acclimation ability of olive plants to adjust to water deficit includes two mechanisms: avoidance and tolerance [3,4]. Among the innate acclimation mechanisms of plants are morphological and physiological leaf alterations; reduction of leaf size and number; biosynthesis and accumulation of compatible solutes (amino acids, proteins, sugars, methylated quaternary ammonium compounds, and organic acids); hormonal balance alteration (abscisic acid-ABA and ethylene); increase in ion efflux with high-affinity antiporters (salt overly sensitive-SOS1 and high-affinity potassium transporters-HKT); maintenance of reactive oxygen species (ROS) homeostasis, and decrease of photosynthetic rates [5,6]. Photosynthetic rates decrease mostly due to stomatal closure, but as water stress becomes severe, the inactivation of photosynthetic activity could be due not only to stomatal restrictions, but also to non-stomatal factors related to inhibition of primary photochemistry and electron transport in chloroplasts [7] as well as to the increase in reactive oxygen species (ROS) levels. When the absorbed light energy is not fully used by photosynthesis, it deviates to molecular oxygen, which is abundant in the chloroplasts [8] leading to ROS formation. ROS are highly reactive oxygen species constantly generated by cell organelles as a metabolic by-product; they function as signaling molecules, but their production is spiked upon stress, and plant normal metabolism is seriously disrupted [9].

Plants have a complex antioxidant system to cope with ROS involving enzymes and molecules [10]. The major enzymatic scavengers of ROS are superoxide dismutase (SOD), ascorbate peroxidase (APX), and catalase (CAT) [11]. In addition, plants contain several low-molecular-weight antioxidants, such as ascorbate, glutathione, and phenolic compounds, which are water-soluble, and tocopherol and carotenoids, which are lipid-soluble [12]. Although studies on the enzymatic antioxidant system of olive trees under water deficit have demonstrated that antioxidant enzymes play a major role in protecting olive leaf tissues against oxidative stress [13–15], the role of phenolic compounds on the water stress tolerance of olive trees has received limited attention [5].

As sessile organisms, olive plants have an active secondary metabolism to improve their adaptation to biotic and abiotic stress conditions [16]. The most characteristic secondary metabolites present in olive trees are iridoids, triterpenes, and phenolic compounds. These metabolites accumulate preferentially in leaves and their beneficial effects as antihypertensive for human health due to the coordinated effects of iridoids (oleuropein, oleacein, and ligustroside) and triterpenes (oleanolic acid) have been previously demonstrated [17–19]; their antitumor potential has also been reported [20].

As water stress is a relevant problem, plants have many innate mechanisms that regulate adaptation to stress. Biotechnological attempts to improve adaptations to water deficit with genetic modifications target the overexpression of different genes, as in *Arabidopsis* [21] or cereals, such as rice [22]. In woody plants like the olive tree with such a large cropping surface, genetic modification is not the best choice to improve adaptation. An alternative to genetic modification addressing several targets is the plants' natural associates, soil microorganisms, especially, beneficial strains termed plant growth-promoting rhizobacteria (PGPR).

The term plant growth-promoting rhizobacteria was coined by Kloepper et al. in 1980 [23] to refer to free-living bacteria that inhabit the rhizosphere, which is the soil closely related to the roots. The mechanisms by which these PGPR improve plant fitness have been largely reviewed [24,25]; PGPR affects plant external factors such as nutrient mobilization or biocontrol of soil microorganisms, or alter internal metabolism by affecting endogenous hormonal balance or systemic induction of metabolism at different levels, like photosynthesis or secondary metabolism. Thus, the role of beneficial rhizobacteria to trigger secondary metabolism appears as a promising alternative to increase the levels of bioactive secondary metabolites [26–29], protect against biotic stress [30] and other frequent situations in agriculture [31]. More precisely, protection against salt stress can be enhanced by beneficial

rhizobacteria by boosting the ROS-scavenging system, increasing compatible solute concentration, such as proline or soluble sugars, or improving photosynthesis and water use efficiency [6].

Therefore, the use of beneficial rhizobacteria capable of modulating secondary metabolism pathways of plants appears as a biotechnological tool with great potential for this purpose. The application of beneficial strains to improve adaptation to abiotic stress has been widely shown for different species, either woody or herbaceous crops, targeting many mechanisms simultaneously [6] and therefore, with great chances of success. To our knowledge, no studies have been undertaken specifically on olive plants to improve adaptation to salt stress with beneficial rhizobacteria, paying specific attention to bioactive molecules accumulated in leaves; furthermore, if bioactives accumulate in leaves, pruning residues can be transformed into a valuable side product, to obtain enriched extracts with antihypertensive potential. Our rationale is that inoculating the olive trees with beneficial rhizobacteria would simultaneously trigger secondary metabolism pathways as well as other mechanisms also involved in abiotic stress adaptation. We selected 10 bacterial strains from a previous screening in *Pinus* rhizosphere [32] to evaluate their ability to improve olive tree adaptation to salt stress and enhance bioactive contents. To achieve this objective, photosynthesis was measured after 12 months of inoculations on plantlets grown in high saline conditions, photosynthetic pigments, osmolites (soluble sugars, proline), ROS scavenging enzymes (SOD, APX), and antioxidant molecules (phenols, flavonols, and oleouropein) were analyzed as markers of the overall fitness of the plant.

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

#### *2.1. Beneficial Strains and Olive Tree Variety*

The 10 beneficial strains (L79, L81, L56, L24, L62, L36, G7, L44, K8, and H47) assayed in this study were isolated from the rhizosphere of *Pinus pinea* and *P. pinaster* [32]. They were able to produce siderophores (L79, L81, G7, H47), auxins (L56, L24, L44), auxins and siderophores (L62, L36) or auxins and degrade ACC (K8). Except for L62, a Gram-positive non-esporulated rod, all other strains are Gram-positive esporulated bacilli.

*Olea europea* (L) var. Arbosana plantlets were used for the study. Plantlets were bought from a commercial producer.

#### *2.2. Inocula Preparation and Delivery to Plants*

Bacterial strains were maintained at −80 ◦C in nutrient broth with 20% glycerol. Inocula were prepared by streaking strains from −80 ◦C onto plate count agar (PCA) plates, incubating plates at 28 ◦C for 24 h. Then, they were grown in Luria Broth liquid media (LB) or nutrient broth (only L62) under shaking (1000 rpm.) at 28 ◦C for 24 h to obtain a 2 <sup>×</sup> 109 cfu/mL inoculum.

These cultures were adjusted to 1 <sup>×</sup> 108 cfu/mL and 500 mL were root-inoculated every 15 days from October 2017 to October 2018.

#### *2.3. Experimental Design*

Six-month olive plantlets were transplanted into 5 L pots with soil from the Guadalquivir Marshes. Plants were arranged in lines on an experimental plot within the marshes (37◦06 34.5 N, 6◦20 22.7 W); pot position was changed every two weeks to avoid side-effects. Plants were watered every 15 days. The electric conductivity of water was 8.20 dS/m and of soil it was 6.07 dS/m.

Bacteria were root-inoculated by soil drench every 15 days from October 2017 to October 2018; so plants received 500 mL of water every week, alternating inoculum and water. Six plants per treatment were inoculated, being one bacterial strain a treatment. Samples were taken in October 2018 and photosynthesis was measured (fluorescence and CO2 fixation). Leaves were powdered in liquid nitrogen and stored at −60 ◦C till analysis. Photosynthetic pigments were determined as well as their antioxidant capacity, analyzing both the enzymatic and non-enzymatic apparatus. Superoxide dismutase (SOD) and ascorbate peroxidase (APX) activities were determined as indicators of the

enzyme apparatus, and total phenols, flavonols and oleuropein, as indicators of the non-enzymatic pool. The osmoprotective effect was evaluated by analyzing compatible solutes (proline and soluble sugars).

#### *2.4. Photosynthesis (Chlorophyll Fluorescence)*

Photosynthetic efficiency was determined through the chlorophyll fluorescence emitted by photosystem II. Chlorophyll fluorescence was measured with a pulse amplitude modulated (PAM) fluorometer (Hansatech FM2, Hansatech, Inc., UK). After dark-adaptation of leaves, the minimal fluorescence (Fo; dark-adapted minimum fluorescence) was measured with a weak modulated irradiation (1 μmol m−<sup>2</sup> s−1). Maximum fluorescence (Fm) was determined for the dark-adapted state by applying a 700 ms saturating flash (9000 μmol m−<sup>2</sup> s<sup>−</sup>1). The variable fluorescence (Fv) was calculated as the difference between the maximum fluorescence (Fm) and the minimum fluorescence (Fo). The maximum photosynthetic efficiency of photosystem II (maximal PSII quantum yield) was calculated as Fv/Fm. Immediately, the leaf was continuously irradiated with red-blue actinic beams (80 μmol m−<sup>2</sup> s<sup>−</sup>1) and equilibrated for 15 s to record Fs (steady-state fluorescence signal). Following this, another saturation flash (9000 μmol m−<sup>2</sup> s<sup>−</sup>1) was applied and then Fm' (maximum fluorescence under light-adapted conditions) was determined. Other fluorescent parameters were calculated as follows: the effective PSII quantum yield ΦPSII = (Fm' − Fs)/Fm' [33]; and the non-photochemical quenching coefficient NPQ = (Fm − Fm')/Fm'. All measurements were carried out in the 6 plants of each treatment.

#### *2.5. Photosynthesis (CO2 Fixation)*

Net photosynthetic rate, (Pn) (mmol CO2/m2), transpiration rate, E (mmol/m2 s) and stomatal conductance, C (mmol/m<sup>2</sup> s) were measured with a portable photosynthetic open-system (CI-340, CID, Camas, WA, USA) [34].

Water use efficiency (WUE) was calculated as net photosynthesis (Pn) divided by transpiration (E) as an indicator of stomatal efficiency to maximize photosynthesis minimizing water loss due to transpiration.

#### *2.6. Photosynthetic Pigments: Chlorophylls and Carotenoids*

Extraction was done according to [35]. One hundred mg of leaves powdered in liquid nitrogen was dissolved in 1 mL of acetone 80% (*v*/*v*), incubated overnight at 4 ◦C and then centrifuged 5 min at 10,000 rpm in a Hermle Z233 M-2 centrifuge. One mL of acetone 80% was added to the supernantant and was mixed with a vortex. Immediately afterward, absorbance at 647, 663, and 470 nm was measured on a Biomate 5 spectrophotometer to calculate chlorophyll a, chlorophyll b, and carotenoids (xanthophylls + carotenes) using the formulas indicated below [35,36].

Chl a (μg/g FW) = [(12.25 × Abs663) − (2.55 × Abs647)] × V(mL)/weight (g).

Chl b (μg/g FW) = [(20.31 × Abs647) − (4.91 × Abs663)] × V(mL)/weight (g).

Carotenoids (μg/g FW) = [((1000 × Abs470) − (1.82 × Chl a) − (85.02 × Chl b))/198] × V(mL)/weight (g).

Tubes were protected from light throughout the whole process.

#### *2.7. Enzymatic Antioxidants: Superoxide Dismutase (SOD) and Ascorbate Peroxidase (APX)*

Before assessing enzymatic activities, soluble proteins were extracted by suspending 100 mg of powder in 1 mL of potassium phosphate buffer 0.1 M, pH 7.0, containing 2 mM phenylmethylsulfonyl fluoride (PMSF). After sonication for 10 min and centrifugation for 10 min at 14,000 rpm, the supernatant was aliquoted, frozen in liquid nitrogen and stored at −80 ◦C for further analysis of APX, SOD, and proteins. All the above operations were carried out at 0–4 ◦C.

To measure the amount of total protein from plant extract, 250 μL of Bradford reagent and 5 μL of sample and BSA dilutions were inoculated in ELISA 96 well plates and incubated for 30 min at room temperature and measured using a plate reader at an absorbance of 595 nm. A calibration curve was constructed from commercial BSA dilutions expressed in milligrams. The protein units were expressed as mg/μL.

APX was measured by the method of Garcia-Limones [37]. The reaction mixture consisted of 50 mM potassium phosphate buffer, pH 7.0, 0.25 mM sodium ascorbate, 5 mM H2O2 and 100 μL of enzyme extract in a final volume of 1.2 mL. Adding H2O2 started the reaction and the oxidation of ascorbate was determined by the decrease in A290. The extinction coefficient of 2.8 mM−<sup>1</sup> cm−<sup>1</sup> was used to calculate activity. One unit of APX activity is defined as the amount of enzyme that oxidizes 1 mmol min−<sup>1</sup> of ascorbate under the above assay conditions.

SOD activity was determined following the specifications of the SOD activity detection kit (SOD Assay Kit-WST, Sigma-Aldrich, Darmstadt, Germany). With this method, xanthine is converted to superoxide radical ions, uric acid, and hydrogen peroxide by xanthine oxidase (XO). Superoxide reacts with WST1 to generate a product that absorbs at around 440 nm. SOD prevents the reduction of WST1 to WST-1formazan, thus reducing the absorption at 440 nm, which is proportional to SOD activity; the rate of the reduction of WST1 with O2 is linearly related to the xanthine oxidase (XO) activity. The unit used for this activity was: % inhibition of WST reduction.

#### *2.8. Osmolites: Proline and Soluble Sugars*

An ethanolic extraction was prepared from 0.25 g of powder and 5 mL of ethanol 70% (*v*/*v*) incubated at 100 ◦C for 20 min.

For proline determination 1 mL of ninhydrin reagent freshly prepared (1 g of ninhydrin dissolved in 60 mL of acetic acid, 20 mL of ethanol and 20 mL of water) was mixed with 0.5 mL of the plant ethanol extract and heated at 95 ◦C for 20 min. Finally, absorbance at 520 nm was measured. Results are expressed as μmol/g.

A soluble sugars determination was performed following Yemm and Willis [38]. Briefly, the following reaction was prepared: 3 mL of the reactive (200 mg of antrone + 100 mL of 72% sulfuric acid) and 0.1 mL of the plant ethanol extract. The reaction was incubated in a bath at 100 ◦C for 10 min. Once it was cold, absorbance was measured at 620 nm. To calculate soluble sugar concentration the following equation was used:

$$\text{mg/g} = \text{I(Abs}\_{620} - 0.016) \text{(0.02)/weight (g)} / 1000 \text{ J}$$

#### *2.9. Total Phenols and Flavonols*

Leaf extracts were prepared from 0.25 g of leaves (powdered in liquid nitrogen) in 2.25 mL methanol 80%, sonicated for 10 min and centrifuged for 5 min at 5000 rpm.

Total phenols were quantitatively determined with Folin-Ciocalteu agent (Sigma. Aldrich, St Louis, MO, USA) by a colorimetric method described by Xu and Chang [39], with some modifications, gallic acid was used as standard (Sigma-Aldrich, St Louis, MO, USA). Twenty microlitres of extract were mixed with 0.250 mL of Folin-Ciocalteu 2 N and 0.75 mL of Na2CO3 20% solution. After 30 min at room temperature, absorbance was measured at 760 nm. A gallic acid calibration curve was made (r = 0.99). Results are expressed in mg of gallic acid equivalents per 100 g of fresh weight (FW).

Total flavonols were quantitatively determined through the test described by Jia et al. [40], using catechin as standard (Sigma-Aldrich, St Louis, MO, USA). One milliliter of the extract was added to a flask of 10 mL with 4 mL of distilled water. After that 0.3 mL of NaNO2 5%, and 0.3 mL of AlCl3 10% were added after 5 min. One minute later, 2 mL of NaOH 1 M were added, and distilled water was added util 10 mL of total volume. The solution was mixed and measured at 510 nm. A catechin calibration curve was made (r = 0.99). Results are expressed as mg of catechin equivalents per 100 g of fresh weigh (FW).

#### *2.10. Oleuropein Extraction and TLC Analysis*

Oleuropein was determined according to the European Pharmacopoeia. One gram of the powdered samples was extracted with 10 mL of methanol under reflux for 15 min. After cooling, samples were filtered and 10 μL was loaded as a band on a TLC silica gel plate; the reference solution contained 10 mg of oleuropein and 1 mg of rutoside trihydrate in 1 mL of methanol. Plates were incubated on a chromatography tank and allowed to develop over a path of 10 cm, being the mobile phase water/methanol/methylene chloride (1.5:15:85 *v*/*v*/*v*). Plates were dried in air. Detection of oleuropein was done by spraying with vanillin sulphuric acid reagent after followed by heating for 5 min at 100–105 ◦C; the brownish-green zone appeared in the middle of the plate was oleuropein and a brownish-yellow zone near the application point was rutoside.

Quantification was done with the image analysis program Quantity One v4.6.8 (Biorad, CA, USA), based on the density and concentration of the oleuropein spot from the reference sample.

#### *2.11. Statistics*

A Principal Components Analysis (PCA) with all the parameters measured for the ten strains was performed with CanocoTM for Windows v.4.5 software (Microcomputer power, Ithaca, USA) [41]. Scaling was performed withinter-species correlation and was achieved dividing by the standard deviation.

To evaluate treatment effects, one way ANOVA (Statpgraphcis Centurion XVIII) were performed for each of the variables. When significant differences appeared (*p* < 0.05), the LSD test (least significant difference) from Fisher was used.

#### **3. Results**

Three main groups can be defined in the ordination provided by principal component analysis (PCA) (Figure 1). The group in the upper part of axis I (dotted line) includes bacteria L36, K8, L44, and L81 and is mainly influenced by photosynthetic pigments concentration and, secondarily, by APX and SOD activities as shown by the length of the vectors. A second group including K8, L44, L81, H47, and L56 (black dashed line) on the left of axis I, can be defined based on phenols and flavonols concentration. A third group formed by L56, L24, and L62 (grey dashed line), is determined by oleuropein concentration and water use efficiency (WUE).

All assayed strains modified photosynthetic parameters (Figure 2), however, Fv/Fm was within normal values (0.82–0.85) and was not significantly affected by any strain. L81, L24, and K8 significantly increased photochemical quenching and all strains except L79 and L36 decreased energy dissipation (NPQ). All of them significantly increased net photosynthesis in terms of CO2 fixation, with an outstanding performance of K8 and L24 that caused 3-fold increases, while all others were in the range of 2-fold increases. Water use efficiency, calculated as the value of net photosynthesis divided by the transpiration rate, was also significantly increased by all strains, with an outstanding performance of L62 that caused a 5-fold increase on WUE, and L24 and L44 in second place, in the range of a 3.5 increase on WUE (Figure 3).

As regards to photosynthetic pigments (Figure 4), controls had around 63 mg/g chlorophyll a, 29.69 chlorophyll b, and 72.3 mg/g carotenes. Most strains maintained or decreased the amount of chlorophylls, except for strain L36 that increased chlorophylls and K8 that increased chlorophyll a; none modified the chlorophyll a/b ratio. The general trend was to lower carotene concentration, except for L36 and L44.

**Figure 1.** Ordination provided by the principal components analysis (PCA) performed with all the parameters measured for the ten strains. APX, ascorbate peroxidase; SOD, superoxide dismutase; NPQ, non-photochemical quenching coefficient; WUE, Water Use Efficiency. Percentages in the axis correspond to the variance absorbed by each of these two first axes.

**Figure 2.** Photosynthetic parameters related to photosystems and light reactions. (**a**) Minimal fluorescence after 20-min dark-adaptation (Fo). (**b**) Maximal PSII quantum yield (Fv/Fm), (**c**) effective PSII quantum yield (ΦPSII) and (**d**) non-photochemical quenching coefficient (NPQ) measured in olive tree plants treated with the ten strains and the non-inoculated control. For each treatment and parameter average value ± standard error value (*n* = 6) is presented. Asterisks (\*) represent significant differences with the control according to the LSD test (*p* < 0.05).

**Figure 3.** Photosynthetic parameters related to C fixation measured in olive tree plants treated with the ten strains. (**a**) Net photosynthesis (PN) measured as the CO2 fixed by the leaves (μmol CO2 /m2 s). (**b**) Water Use Efficiency (WUE) calculated as PN divided by transpiration rate (μmolH2O /m<sup>2</sup> s). Average values of the replicates with standard error bars are represented (*n* = 6). Asterisks (\*) represent significant differences with the control according to the LSD test (*p* < 0.05).

**Figure 4.** Photosynthetic pigments concentration (μg/g FW). (**a**) Chlorophyll a, (**b**) Chlorophyll b and (**c**) Carotenoids were measured in olive tree leaves treated with the ten strains. For each treatment and parameter average value ± standard error value is presented (*n* = 6). Asterisks (\*) represent significant differences with the control according to the LSD test (*p* < 0.05).

As for the enzymatic antioxidant systems (Figure 5), only L81 increased the activity of SOD and L44 that of APX, while no changes or slight decreases were induced by the other 8 strains.

**Figure 5.** Enzyme activities related to oxidative stress. (**a**) Superoxide dismutase (SOD), (**b**) Ascorbate peroxidase (APX) activities measured in olive tree leaves treated with the ten strains. Enzymatic activities were calculated as mmol/mg protein min (for APX) and % of inhibition/mg protein (for SOD). For each treatment and parameter average value ± standard error value is presented (*n* = 6). There are no significant differences according to the LSD test (*p* < 0.05).

Soluble sugars in controls were 3.77 mg/g and proline contents 0.37 μmol/g (Figure 6). Soluble sugars contents were significantly increased by all treatments, ranging from 10% (L24) to 30% (L62 (Figure 6a)). Similarly to soluble sugars, proline was significantly increased by most strains ranging from 20% (L56) to 50% (G7, H47); only L81 and L62 showed similar proline contents than controls (Figure 6b).

**Figure 6.** (**a**) Proline (nmol/g fresh weight) and (**b**) soluble sugars concentration (μmol/g fresh weight) measured in olive tree leaves treated with the ten strains. Average values of the replicates with standard error bars are represented (*n* = 6). Asterisks (\*) represent significant differences with the control according to the LSD test (*p* < 0.05).

As for bioactives (Figure 7), controls had high phenols concentration (1295 meq gallic acid/100 g FW) and low flavonols (6.46 meq catechin/100 g FW), and 9.11 mg oleuropein/g. Strains L81 and L56 increased the concentration of phenols and flavonols in leaves and L44, K8, and H47 only that of phenols. Only L81 increased oleuropein concentration (12%), L62 did not affect it and most strains decreased the amount of oleuropein, a group with minor decreases (L56, L24, L79) and another group with major decrease.

**Figure 7.** Bioactives. (**a**) Phenols (meq gallic acid/100 g fresh weight), (**b**) flavonols (meq catechin/100 g fresh weight) and (**c**) oleuropein (mg/mL) concentration measured in olive tree leaves treated with the ten strains. For each treatment and parameter average value ± standard error value is presented. There are no significant differences according to the LSD test (*p* < 0.05).

#### **4. Discussion**

In this study, the ability of the ten beneficial rhizobacteria assayed to modify plant physiology of olive plantlets growing in pots with high electric conductivity of soil and water when root-inoculated has been evidenced. According to Chartzouaskis [42], values of water (8.20 dS/m) and soil (6.07 dS/m) electric conductivity, indicate severe salinity for olive trees in the present study. In these harsh conditions, all strains are able to trigger plant adaptative metabolism, improving net photosynthesis and mainly affecting osmolite concentration. The positive effect of these strains was expected since the original screening rendered several beneficial strains, two of which (L62 and L81) are also tested in this experiment [30,32].

In general, the effects of bacteria on plant metabolism and physiology target photosynthetic pigments, photosynthesis, and osmolites [8] as revealed by the principal components analysis (PCA), a multivariate analysis that provides an ordination of the samples based on their similarity considering all the variables analyzed. In this ordination (Figure 1), samples are grouped mainly due to the differential effects of bacteria in those three variables. The group in the upper part of axis I (dotted line) includes bacteria that increase (L36 and K8) or maintain (L44 and L81) photosynthetic pigments. The position of these two bacteria (L44 and L81) in the ordination is also determined by the effect they have in APX and SOD, respectively. All the other bacteria decreased pigment concentration. Bacteria in the second group (black dashed line), K8, L44, L81, H47, and L56, increase leaf phenol concentrations, and L81 and L56 also increase flavonols. A third group formed by L56, L24, and L62 (grey dashed line) includes the three bacteria that maintain oleuropein concentration similar to control plants, while the other bacteria reduce this concentration, except for L81, the only strain that increases

oleuropein concentration in leaves. Finally, L24 and L62 induced the highest water use efficiency in the plants. Therefore, all these parameters are differentially affected by the different strains while proline, soluble sugars concentration, and net photosynthetic rate are similarly triggered by all bacteria [6], as can be noticed by the shorter length of vectors.

Under mild and moderate water stress, photosynthetic rate decreases in olive plants mostly due to stomatal closure [43]. However, as water stress becomes severe, the inactivation of photosynthetic activity could be due not only to stomatal closure but also to non-stomatal factors related to inhibition of primary photochemistry and electron transport in chloroplasts [7]. A decrease in chlorophylls under salt stress has been explained by pigment destruction after the ROS peak [9]. However, if this was the case, photosynthesis would be impaired due to cell membrane alterations and lack of pigments. Interestingly, our data show increased effective PSII quantum yield (ΦPSII) (L81, L24 and K8), and lower NPQ, suggesting that bacteria are triggering an innate plant protective mechanism against the excess of light entering the system, and making better use of the energy fixed.

Under salt stress conditions olive leaves become thicker [44] compromising CO2 diffusion to chloroplasts [45,46]. Photosynthesis is reduced under saline stress in olive trees [44,47,48], but with different effects on the CO2 assimilation rate depending on salt concentration. Interestingly, all strains increased net photosynthesis, consistent with the reported modification of carbohydrate production and sink utilization that leads to downregulate feedback photoinhibition and boost plant energy metabolism [49], probably providing C scaffoldings for secondary metabolites and osmolyte synthesis and accumulation. Despite the significant increase in C fixation induced by all strains, water use efficiency (WUE) was different, with a striking two-fold increase for most strains except for L62, which showed a 4-fold increase, and L44 and L24, showing a 3-fold increase (Figure 3). This indicates a strong improvement in plant fitness in a high saline environment, suggesting activation of protective systems and highlighting the different mechanisms involved in adaptation to harsh conditions and supporting the multivariate solution provided by PGPR [6].

There is a demonstrated relationship between compatible solutes and photosynthesis. All strains increased concentration of compatible solutes as a protective mechanism, since they sequester water molecules, protect cell membranes and protein complexes, and allow the metabolic machinery to continue functioning [8]. Consistent with our data, carbohydrates are the most common solutes accumulated in olive tree tissue under water deficit conditions [6,50], and all strains significantly increased them, being thus a primary defense mechanism [49]. The close relationship between net photosynthetic rate and proline content reported by BenAhmed et al. [13], is consistent with our data, as proline was increased by all except for L81 and L62, confirming the important role of this osmolyte in the maintenance of photosynthetic activity and plant homeostasis. The different behaviors suggest the involvement of other factors, as L81 and L62 performed outstandingly.

Bacterial strains modify differently innate plant mechanisms to cope with ROS [29,51]; while the enzyme ROS scavenging system is hardly modified by bacteria, phenolic compounds are. The enzyme system is probably in its maximum natural activation, which cannot be further increased by bacteria; however, L81 is still able to significantly enhance SOD activity. Although studies on the enzymatic antioxidant system of the olive tree under water deficit have demonstrated that antioxidant enzymes play a major role in protecting olive leaf tissue against oxidative stress [13–15], limited attention has been given to the effect of phenolic compounds on water stress tolerance. Phenolic compounds are constitutively present in all higher plants. However, phenylpropanoid metabolism is often induced when plants are exposed to a wide range of environmental stresses [52], including bacteria [26,53]. In view of our results, abiotic stress as well as beneficial rhizobacteria modified the antioxidant pools of the plants but in an uncoupled way. Some of the assayed strains increased total phenols and flavonols (L81 and K8), and oleuropein (only L81), while another group decreased concentrations of these metabolites (L79, L24, L62, and G7). Among the first group, L81 increased SOD enzyme activity. From the second group, L79 increased SOD and L24 decreased both SOD and APX enzymatic activities, which suggests either higher oxidative stress as a consequence or other mechanisms to cope with

ROS (Figure 5). The different influence of bacterial strains on ROS scavenging enzyme activity has been described before to be strain-specific [28]. García-Cristobal et al [28] reported the ability of a *Chryseobacterium* strain to enhance ROS scavenging enzymes activity in salt and pathogen stressed rice plants, while a *Pseudomonas* strain enhanced protection by increasing defensive enzymes, not ROS scavenging enzymes.

Under salt stress conditions, phenolic compounds produced in leaves increase [5,6,15]. However, in this work only five strains increased total phenolics concentration and only two significantly increased total flavonols, while others decreased them, reinforcing the species specificity between plant and bacteria [28,53], a receptor-mediated effect, hence highly specific. Irrespective of the final phenolics balance, all bacteria have altered this pathway, confirming the role of this pathway in adaptation; not only phenolics behave as antioxidants, but other derived molecules may also have this role [27,54]. Considering all this data together, it seems that bacteria are lowering photosynthetic pigments concentration, suggesting this effect as a mechanism to decrease oxidative stress due to photosynthesis, especially since the enzymatic antioxidant pool is not affected or even decreased, and bioactives are only increased by half of the strains. Finally, oleuropein, a bioactive molecule accumulated in leaves [18], with a proposed role as a protective molecule against biotic stress due to its potential as a cross-linking agent [54], is only increased under the influence of L81. Again, two strategies are depicted, either slightly lowering its concentration, or minimizing it reinforcing the hypothesis of each strain activating different mechanisms of plant adaptation.

Not only olive oil is obtained from olive trees; also solid residues obtained in considerable amounts during olive oil production and elaboration of table olives are of great concern in the Mediterranean area, as these by-products accumulate in large amounts. Great progress has been made to recycle these residues obtaining an economic profit, like obtaining activated carbon [55–58] or fuel for the generation of heat and electricity [59–61]. However, olive leaves, which are produced in large amounts, render scarce profits at present. Nevertheless, the market for natural ingredients and additives is rapidly growing, with such products obtaining high prices. Increased concentrations of phenolic compounds and especially oleuropein, with strong antihypertensive potential, reinforces the potential of olive leaves in the field of a circular economy. Furthermore, delivering beneficial strains to edible plants has improved beneficial effects for health as not only the targeted metabolites are increased, but there is a general physiological change that results in improved effects on health [53].

In summary, delivering beneficial strains improves adaptation to high saline conditions, mainly affecting osmolytes and improving net photosynthesis and water use efficiency. Interestingly, L81 differentially increases oleuropein constituting a good treatment to improve the potential of olive leaves for its antihypertensive effects. L62 is the one to improve WUE, which is especially good to improve plant adaptation to harsh conditions of low water availability.

#### **5. Conclusions**

In view of these data, it is evidenced that all bacterial strains improve plant adaptation increasing osmoprotection and net photosynthesis but they differentially affect the enzymatic and non-enzymatic antioxidant systems. Bacteria able to increase bioactive concentration and therefore potential benefits of olive leaves on health may also contribute to a circular economy, recycling pruning residues.

**Author Contributions:** Conceptualization, F.J.G.-M., B.R.-S., A.G.-V.; methodology, A.G.-V., E.G.-C.; formal analysis, E.G.-C., M.B.M.-P.; investigation, E.G.-C.; resources, F.J.G.-M., B.R.-S.; data curation, A.G.-V., E.G.-C.; writing—original draft preparation, A.G.-V.; writing—review and editing, A.G.-V., B.R.-S.; supervision, F.J.G.-M., B.R.-S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Authors would like to thank J.C. for providing experimental fields and local help.

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

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

### *Article* **New Evidences of Antibacterial E**ff**ects of Cranberry Against Periodontal Pathogens**

#### **María C. Sánchez 1, Honorato Ribeiro-Vidal 1, Begoña Bartolomé 2, Elena Figuero 1, M. Victoria Moreno-Arribas 2, Mariano Sanz <sup>1</sup> and David Herrera 1,\***


Received: 20 January 2020; Accepted: 20 February 2020; Published: 24 February 2020

**Abstract:** The worrying rise in antibiotic resistances emphasizes the need to seek new approaches for treating and preventing periodontal diseases. The purpose of this study was to evaluate the antibacterial and anti-biofilm activity of cranberry in a validated in vitro biofilm model. After chemical characterization of a selected phenolic-rich cranberry extract, its values for minimum inhibitory concentration and minimum bactericidal concentration were calculated for the six bacteria forming the biofilm (*Streptococcus oralis*, *Actinomyces naeslundii*, *Veillonella parvula*, *Fusobacterium nucleatum*, *Porphyromonas gingivalis*, and *Aggregatibacter actinomycetemcomitans*). Antibacterial activity of the cranberry extract in the formed biofilm was evaluated by assessing the reduction in bacteria viability, using quantitative polymerase chain reaction (qPCR) combined with propidium monoazide (PMA), and by confocal laser scanning microscopy (CLSM), and anti-biofilm activity by studying the inhibition of the incorporation of different bacteria species in biofilms formed in the presence of the cranberry extract, using qPCR and CLSM. In planktonic state, bacteria viability was significantly reduced by cranberry (*p* < 0.05). When growing in biofilms, a significant effect was observed against initial and early colonizers (*S. oralis* (*p* ≤ 0.017), *A. naeslundii* (*p* = 0.006) and *V. parvula* (*p* = 0.010)) after 30 or 60 s of exposure, while no significant effects were detected against periodontal pathogens (*F. nucleatum*, *P. gingivalis* or *A. actinomycetemcomitans* (*p* > 0.05)). Conversely, cranberry significantly (*p* < 0.001 in all cases) interfered with the incorporation of five of the six bacteria species during the development of 6 h-biofilms, including *P. gingivalis*, *A. actinomycetemcomitans*, and *F. nucleatum*. It was concluded that cranberry had a moderate antibacterial effect against periodontal pathogens in biofilms, but relevant anti-biofilm properties, by affecting bacteria adhesion in the first 6 h of development of biofilms.

**Keywords:** polyphenols; cranberry; periodontal diseases; dental biofilm; antibacterial activity; anti-biofilm activity; *F. nucleatum*; *P. gingivalis*; *A. actinomycetemcomitans*

#### **1. Introduction**

Dental biofilm-organized periodontal pathogens (including *Porphyromonas gingivalis* and *Aggregatibacter actinomycetemcomitans*) are the primary etiological factor of periodontal diseases, which are one of the most prevalent conditions affecting human beings [1]. These conditions have not only a relevant impact in the mouth [1], but also in systemic health [2] and in quality of life indicators [3]. Due to the infectious nature of periodontal diseases, antimicrobials are widely used in their management (prevention and treatment) [4,5]. However, the worrying rise in antibiotic resistances, including those in periodontal pathogens [6] and unwanted effects of antiseptics/antimicrobials compounds [4,5] emphasize the need to seek new approaches for treating and preventing periodontal

diseases. Therefore, attention is given to the need of finding, developing, and improving antimicrobial natural compounds, capable of inhibiting the proliferation and/or adhesion of bacteria pathogens in dental/oral biofilms [5,7–10].

In previous studies, it has been shown that polyphenols, and other compounds derived from plants have an influence on human microbiota, either by promoting the growth of beneficial microorganisms or by acting against pathogens [11,12]. Cranberry (*Vaccinium macrocarpum*) compounds, including phenolic acids, proanthocyanidins (particularly, A-type proanthocyanidins), anthocyanins, organic acids, and their microbial-derived metabolites [13], selectively inhibit the growth of intestinal pathogens such as *Staphylococcus strains* and *Salmonella enterica* [14], reduce *Escherichia coli* colonization of the urinary tract [15–17], restrict the virulence of *Pseudomonas aeruginosa* [18,19], present anti-oxidant potential [20], anti-adhesion of Gram-negative and Gram-positive bacteria [21,22], and anti-motility [23,24]. Furthermore, they may be associated with relevant health benefits, including a decreased risk of cardiovascular disease-related mortality [25], prevention of type 2 diabetes mellitus [26], and potential anti-cancer properties [27,28].

The antibacterial and anti-adhesion features of cranberry against oral bacteria have drawn wide attention [22,29–31]. Several in vivo and in vitro studies have evaluated how certain cranberry derived compounds could interfere with formation of a cariogenic biofilm. In this regard, it has been demonstrated that certain components of cranberries may limit dental caries by inhibiting the production of organic acids by cariogenic bacteria, the formation of biofilms by *Streptococcus mutans* and *Streptococcus sobrinus*, and the adhesion and coaggregation of a considerable number of other oral species of *Streptococcus* [32–36]. Focusing on periodontal diseases, the non-dialyzable constituent fraction of cranberry (NDM) inhibits the formation of *P. gingivalis* [37] and *Fusobacterium nucleatum* [38] biofilms, two bacteria species associated with periodontitis. The NDM fraction may also inhibit the adhesion of *P. gingivalis* to various proteins, including type I collagen [37] and may reduce bacterial coaggregation involving periodontal pathogens [32]. However, the information on the antibacterial and anti-biofilm capacity of natural extracts from cranberry against relevant periodontal pathogens, growing in complex multi-species biofilms, is scarce.

Therefore, the aim of the present study was to evaluate the antibacterial and anti-biofilm activity of cranberry extracts in a multispecies in vitro biofilm model, including six bacteria species (*Streptococcus oralis*, *Veillonella parvula*, *Actinomyces naeslundii* and the periodontal pathogens *P. gingivalis*, *A. actinomycetemcomitans*, and *F. nucleatum*). The specific objectives were to assess (1) the antibacterial activity of a cranberry extract against bacteria species in formed biofilms, by assessing the reduction in bacteria viability, and (2) the anti-biofilm activity, by studying the inhibition of the incorporation of different bacteria species in biofilms formed in the presence of the cranberry extract.

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

#### *2.1. Cranberry Extract*

The cranberry extract used in this study was provided by Triarco Industries Inc. (Cranbury, NJ, USA). For determination of its total polyphenols content, the extract (0.05 g) was dissolved in 10 mL of methanol/HCl (1000:1, *v*/*v*), sonicated (120 W) for 5 min followed by an extra 15 min resting period, centrifuged, and filtered through a 0.22-μm membrane filter. For analysis of individual phenolic compounds, the extract (0.50 g) was dissolved in 10 mL of MeOH/H2O (20:80, *v*/*v*) containing 0.2% HCl, sonicated for 10 min, centrifuged, and filtered through 0.22 μm. In both cases, sample preparations were performed in duplicate.

#### *2.2. Analysis of Phenolic Compounds in the Cranberry Extract*

Total polyphenols content was measured by the Folin-Ciocalteu reagent (Merck, Darmstadt, Germany) and using gallic acid (25–500 mg L−1) as a calibration standard. Analysis of individual phenolic compounds was carried out by UPLC-DAD-ESI-TQ MS, as previously described in

Sanchez-Patán et al. [39]. Different phenolic acids (including phenylpropionic, phenylacetic, mandelic, benzoic, and cinnamic acids), flavan-3-ols (monomers, B-type procyanidin dimers and trimers, and A-type procyanidin dimers and trimers), and anthocyanins (peonidin, cyanidin, and malvidin derivatives) were targeted [39]. Commercial standards of these phenolic acids were used to construct calibration curves for sample quantification [39].

#### *2.3. Bacteria Strains and Culture Conditions*

Reference strains of *S. oralis* CECT 907T, *V. parvula* NCTC 11810, *A. naeslundii* ATCC 19039, *F. nucleatum* DMSZ 20482, *A. actinomycetemcomitans* DSMZ 8324, and *P. gingivalis* ATCC 33277 were used. These bacteria were grown on blood agar plates (Blood Agar Oxoid No 2; Oxoid, Basingstoke, UK), supplemented with 5% (*v*/*v*) sterile horse blood (Oxoid), 5.0 mg L−<sup>1</sup> hemin (Sigma, St. Louis, MO, USA) and 1.0 mg L−<sup>1</sup> menadione (Merck, Darmstadt, Germany) in anaerobic conditions (10% H2, 10% CO2, and balance N2) at 37 ◦C for 24–72 h.

#### *2.4. Antibacterial Assays*

Figure 1 shows the experimental design followed for the study of the antibacterial effects of cranberry against planktonic bacteria and in an oral biofilm model.

**Figure 1.** Scheme of the antibacterial assays carried out in this study.

#### 2.4.1. Antibacterial Effect of Cranberry Extract Against Planktonic Bacteria

Pure cultures of the bacteria species were grown anaerobically in a protein rich medium containing brain-heart infusion (BHI) (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) supplemented with 2.5 g L−<sup>1</sup> mucin (Oxoid), 1.0 g L−<sup>1</sup> yeast extract (Oxoid), 0.1 g L−<sup>1</sup> cysteine (Sigma), 2.0 g L−<sup>1</sup> sodium bicarbonate (Merck), 5.0 mg L−<sup>1</sup> hemin (Sigma), 1.0 mg L−<sup>1</sup> menadione (Merck), and 0.25% (*v*/*v*) glutamic acid (Sigma). The bacteria growth was harvested at mid-exponential phase (measured by spectrophotometry). Microtitre plate-based antibacterial assays were carried out in a 96-wells plate, adding 190 μL of each bacteria inoculum at a final concentration of 10<sup>6</sup> colony forming units (CFUs) mL−1, and 10 μL of the sterile cranberry extract at a final concentration of 1.0, 0.50, 0.25, 0.10, and 0.01 mg mL−1. Plates had a set of controls: negative control (culture media without any inoculum/cranberry extract), positive control (bacteria without any treatment) as well as blanks

(cranberry extract or dimethyl sulfoxide (DMSO) dissolved in the culture media), to ensure the validity of the assay, 4% DMSO (to identify a possible bactericidal effect of DMSO, used as a solvent for the cranberry extract), and 0.2% chlorhexidine (CHX), in order to compare with the reference of known antibacterial effect. A measurement (optical density, O.D.595) as t = 0 absorbance was taken in a microtitre plate reader (Optic Ivymen System 2100-C; I.C.T.; La Rioja, Spain). The microplates were incubated for 48 h at 37 ◦C under anaerobic conditions, and absorbance was measured at selected intervals (1 h during the first 12 h, and every 12 h to complete 48 h), in order to determine the bacteria growth in time, until the bacteria reached stationary growth phase. MIC (minimum inhibitory concentration) and MBC (minimum bactericidal concentration) values were calculated and confirmed by microbial plate counting on blood agar media. Accordingly, the lowest concentration of the cranberry extract showing growth inhibition was considered as the MIC, whereas the lowest concentration of the cranberry extract that showed zero growth in blood agar plates, after spot inoculation and incubation for 72 h, was recorded as the MBC. All experiments were performed in triplicate with appropriate controls.

#### 2.4.2. Antibacterial Effect in an Oral Biofilm Model in Vitro

In order to optimize the method for evaluating the antibacterial effect of the cranberry extract against the bacteria species growing in biofilms, a range of cranberry concentrations were initially tested (from MBCs to stock solution of cranberry extracts at 20 mg mL−1). A dose of 20 mg mL−<sup>1</sup> yielded the higher antibacterial effect (data not shown).

The multi-species in vitro biofilm model was developed as previously described by Sánchez et al. [40]. Briefly, pure cultures of each bacteria specie were grown anaerobically in the supplemented BHI medium. The bacteria growth was harvested at mid-exponential phase (measured by spectrophotometry), and a mixed bacteria suspension in modified BHI medium containing 103 CFUs mL−<sup>1</sup> for *S. oralis*, 10<sup>5</sup> CFUs mL−<sup>1</sup> for *V. parvula* and *A. naeslundii*, and 106 CFUs mL−<sup>1</sup> for *F. nucleatum*, *A. actinomycetemcomitans* and *P. gingivalis* was prepared (different concentrations based on the different growth rates of each bacteria species). Sterile calcium hydroxyapatite (HA) discs of 7 mm of diameter and 1.8 mm (standard deviation, SD = 0.2) of thickness (Clarkson Chromatography Products, Williamsport, PA, USA) were coated with sterile saliva for 4 h at 37 ◦C in sterile plastic tubes to allow the formation of the acquired pellicle [40], and then placed in the wells of a 24-well tissue culture plate (Greiner Bio-one, Frickenhausen, Germany). Each well was inoculated with 1.5 mL mixed bacteria suspension prepared and incubated in anaerobic conditions (10% H2, 10% CO2, and balance N2) at 37 ◦C for 72 h.

After 72 h, biofilms were dipped during 30 s and 60 s in the cranberry solution (20 mg mL−1), at room temperature. Exposure time of 30 and 60 s were selected since cranberry extracts are bioactive products, commercially available, and for them, the standard exposure times established for other antimicrobial commercially available products (e.g., chlorhexidine mouth rinses), were selected [41–43]. Phosphate buffered saline solution (PBS) was used as negative control and, in order to discard a bactericidal effect of DMSO used to dissolve the extracts, a 4% DMSO solution was also tested.

The antibacterial activity in 72 h biofilms was examined by determining the reduction in the number of viable bacteria counts (expressed as CFUs mL−1), using quantitative polymerase chain reaction (qPCR) combined with Propidium Monoazide (PMA), and by Confocal Laser Scanning Microscopy (CLSM). Assays were conducted in triplicate (with trios of biofilms per replica).

#### *2.5. Anti-Biofilm Assay*

In order to optimize the method for evaluating the anti-biofilm effect of cranberry extracts against the selected bacteria species, different concentrations were tested, based on MICs of each bacteria species in planktonic state (data not shown), and it was finally concluded that a dose of 0.20 mg mL−<sup>1</sup> provided the largest anti-biofilm impact, without affecting bacteria viability in planktonic state.

For the anti-biofilm assay, the mixed bacteria suspension in modified BHI medium containing 103 CFUs mL−<sup>1</sup> for *S. oralis*, 105 CFUs mL−<sup>1</sup> for *V. parvula* and *A. naeslundii*, and 10<sup>6</sup> CFUs mL−<sup>1</sup> for *F. nucleatum*, *A. actinomycetemcomitans* and *P. gingivalis* was prepared as previously described. HA discs were coated with treated saliva for 4 h at 37 ◦C in sterile plastic tubes, and then placed in the wells of a 24-well tissue culture plates. Each well was inoculated with 1.5 mL mixed bacteria suspension prepared and the cranberry extract at 0.20 mg mL−1, or with PBS and DMSO in control biofilms, were added. Plates were incubated in anaerobic conditions, at 37 ◦C for 6 h.

The anti-biofilm activity was examined by determining bacteria counts in biofilms, as CFUs mL−<sup>1</sup> by means of qPCR, and by CLSM. Assays were conducted in triplicate (with trios of biofilms per replica).

#### *2.6. Microbiological Outcomes*

After antibacterial and anti-biofilm assays, biofilms were recovered and sequentially rinsed in 2 mL of sterile PBS (immersion time per rinse, 10 s) three times, in order to remove possible remnants of the extracts and non-adherent bacteria. Then, biofilms were disrupted by vortex for 2 min in 1 mL of PBS. In the case of antibacterial activity, and to discriminate between DNA from live and dead bacteria, PMA was used (Biotium Inc., Hayword, CA, USA). The use of this PMA dye has shown the ability to distinguish between viable and irreversibly damaged cells and hence when combined with qPCR to detect the DNA from viable bacteria [44]. PMA was added to sample tubes containing 250 μL of disaggregated biofilm cells, at a final concentration of 100 μM. Following an incubation period of 10 min at 4 ◦C in the dark, the samples were subjected to light-exposure for 30 min, using PMA-Lite LED Photolysis Device (Biotium Inc.). After PMA photo-induced DNA cross-linking, the cells were centrifuged at 15,000 rcf for 3 min prior to DNA isolation.

Bacteria DNA was isolated from all biofilms using a commercial kit ATP Genomic DNA Mini Kit® (ATP biotech. Taipei, Taiwan), following manufacturer's instructions and the hydrolysis 5'nuclease probe assay qPCR method was used for detecting and quantifying the bacteria DNA. The qPCR amplification was performed following a protocol previously optimized by our research group, using primers and probes targeted against 16S rRNA gene (obtained through Life Technologies Invitrogen (Carlsbad, CA, USA) and Applied Biosystems (Carlsbad, CA, USA)) [44].

Each DNA sample was analyzed in duplicate. Quantification cycle (Cq) values, previously known as cycle threshold (Ct) values, describing the PCR cycle number at which fluorescence rises above the baseline, were determined using the provided software package (LC 480 Software 1.5; Roche Diagnostic GmbH; Mannheim, Germany). Quantification of viable cells by qPCR was based on standard curves. The correlation between Cq values and CFUs mL−<sup>1</sup> was automatically generated through the software (LC 480 Software 1.5; Roche).

All assays were developed with a linear quantitative detection range established by the slope range of 3.3–3.7 cycles/log decade, r<sup>2</sup> > 0.998, and an efficiency range of 1.9–2.0.

#### *2.7. Confocal Laser Scanning Microscopy (CLSM) Analyses*

After antibacterial and anti-biofilm treatment referred above, and before CLSM analysis, the discs were sequentially rinsed in 2 mL of sterile PBS (immersion time per rinse, 10 sec) three times, in order to remove possible remnants of the extract and non-adherent bacteria. Non-invasive confocal imaging of fully hydrated biofilms was carried out using a fixed-stage Ix83 Olympus inverted microscope coupled to an Olympus FV1200 confocal system (Olympus; Shinjuku, Tokyo, Japan). Specimens were stained with LIVE/DEAD® BacLightTM Bacteria Viability Kit solution (Molecular Probes B. V., Leiden, The Netherlands) at room temperature. The 1:1 fluorocrome ratio with a staining time of 9 ± 1 min was used to obtain the optimum fluorescence signal at the corresponding wave lengths (Syto9: 515–530 nm; Propidium Iodide (PI): >600 nm). At least three separate and representative locations on the discs covered with biofilm were selected for these measurements (based on the presence of stacks or "towers" identified in the confocal field view). The CLSM software was set to take a z-series

of scans (xyz) of 1 μm thickness (8 bits, 1024 × 1024 pixels). Image stacks were analyzed by using the Olympus® software (Olympus). Image analysis and live/dead cell ratio (i.e., the area occupied by living cells divided by the area occupied by dead cells) was performed with Fiji software (ImageJ Version 2.0.0-rc-65/1.52b, Open source image processing software).

#### *2.8. Statistical Analyses*

The selected outcome variables to study the antibacterial effect of cranberry extracts were the counts of viable bacteria present on the biofilms, expressed as viable CFUs mL−<sup>1</sup> of *S. oralis*, *V. parvula*, *A. naeslundii*, *F. nucleatum*, *A. actinomycetemcomitans*, and *P. gingivalis* by qPCR, and the live/dead cell ratio of the whole biofilm by CLSM. An experiment-level analysis was performed for each parameter of the study (n = 9 for qPCR and n = 3 for CLSM results). Shapiro–Wilk goodness-of-fit tests and distribution of data were used to assess normality. The effect of each solution (cranberry extracts, PBS and 4% DMSO), the time of exposure (30 or 60 s), and their interaction with the main outcome variable (counts expressed as CFUs mL−<sup>1</sup> or live/dead cell ratio), was compared by means of a parametric ANOVA test for independent samples, and a general linear model was constructed for each bacterium for qPCR results and for total bacteria for live/dead cell ratio of whole biofilm obtained by CLSM, using the method of maximum likelihood and Bonferroni corrections for multiple comparisons.

To study the anti-biofilm effect of the cranberry extract, the selected outcome variables were the counts of bacteria present on the biofilms, expressed as CFUs mL−<sup>1</sup> of *S. oralis*, *V. parvula*, *A. naeslundii*, *F. nucleatum*, *A. actinomycetemcomitans*, and *P. gingivalis* by qPCR, and the live/dead cell ratio of the whole biofilm by CLSM. Shapiro–Wilk goodness-of-fit tests and distribution of data were used to assess normality. An experiment-level analysis was performed for each parameter of the study (n = 9 for qPCR and n = 3 for CLSM results). The effect of each solution (cranberry extract, PBS, and 4% DMSO) on the main outcome variables (CFUs mL−<sup>1</sup> or live/dead cell ratio), was compared by means of a parametric ANOVA test for independent samples, using the method of maximum likelihood and Bonferroni corrections for multiple comparisons.

Data was expressed as means ± SD and as the mean percent inhibition that was calculated by Equation: Percent inhibition = (CFUs mL−<sup>1</sup> of negative control–CFUs mL−<sup>1</sup> of test / CFUs mL−<sup>1</sup> of negative control) × 100.

Results were considered statistically significant at *p* < 0.05. A software package (IBM SPSS Statistics 24.0; IBM Corporation, Armonk, NY, USA) was used for all data analysis.

#### **3. Results**

#### *3.1. Phenolic Composition of the Cranberry Extract*

A phenolic characterization of the cranberry extract was initially carried out to ensure its susceptibility for this study. The content in total polyphenols content of the extract resulted in 219 mg of gallic acid equivalents g−1. Concerning phenolic composition, Table 1 details the main phenolic compounds found in the extract, as determined by UPLC-DAD-ESI-TQ MS. Among phenolic acids (benzoic and cinnamic acids), the extract was especially rich in benzoic acid (8.38 mg g<sup>−</sup>1), followed by others such as p-coumaric acid (0.84 mg g−1) and protocatechuic acid (0.73 mg g−1) in considerable less content. Concerning flavan-3-ols, main MS signals corresponded to A-type trimers (1.58 mg g<sup>−</sup>1), followed by A-type (0.23 mg g−1) and B-type (0.20 mg g−1) dimers, monomers (0.065 mg g−1), and B-type trimers (0.034 mg g<sup>−</sup>1). In relation to anthocyanins, peonidin-3-arabinoside (0.32 mg g−1) and cyanidin-3-arabinoside (0.15 mg g<sup>−</sup>1) showed the highest content (Table 1). These compositional data were in accordance to others commercial cranberry extracts [39].


**Table 1.** Phenolic compounds present in the cranberry extract used in this study. Data are expressed as mean and standard deviation (SD).

#### *3.2. Antibacterial Assays*

#### 3.2.1. Antibacterial Effect of Cranberry Extract Against Planktonic Bacteria

MICs and MBCs values against the six bacteria species selected in planktonic state were determined for the selected cranberry extract. MICs indicated an average bacteriostatic concentration of 0.10 mg mL−<sup>1</sup> against *P. gingivalis* and *F. nucleatum*, 0.25 mg mL−<sup>1</sup> for *A. naeslundii* and *A. actinomycetemcomitans*, 0.50 mg mL−<sup>1</sup> for *V. parvula*, and >1.00 mg mL−<sup>1</sup> for *S. oralis*. MBCs tests showed similar results, with bactericidal concentrations of 0.25 mg mL−<sup>1</sup> against *P. gingivalis*, 1.00 mg mL−<sup>1</sup> against *F. nucleatum*, and >1.00 mg mL−<sup>1</sup> for *S. oralis, A. naeslundii, V. parvula,* and *A. actinomycetemcomitans*. According to these results, the cranberry extract exhibited antibacterial activity, displaying the largest antibacterial properties against the periodontal pathogens *P. gingivalis* and *F. nucleatum.*

3.2.2. Antibacterial Effects in an in Vitro Biofilm Model: Bacteria Counts

Table 2 depicts the effect of cranberry extracts (20 mg mL−1), compared to the negative control (PBS) and 4% DMSO control solution, on the mean number of viable bacteria counts in 72 h biofilms. After an exposure of 30 or 60 s to the cranberry extract, significant reductions in viable counts in biofilms were observed for initial and early colonizers. *S. oralis* showed significant reductions after 30 (*p* < 0.001) and 60 s (*p* = 0.017) when compared to negative control (PBS), reaching in both cases a decrease of 98.9% of viable CFUs (Table 2). Significant differences (*p* < 0.001 after 30 s of exposure) were also observed when the effects of DMSO solution was compared to PBS, with percentages of decrease of 93.1% and 58.8% for 30 and 60 s exposures, respectively (Table 2). For *A. naeslundii* and *V. parvula*, a significant impact of the cranberry extract was observed after 30 s (65.7% of reduction, *p* = 0.006 and 66.7% of reduction, *p* = 0.010, respectively), but not after 60 s. No significant reductions were observed after exposure to DMSO (*p* > 0.05), after 30 or 60 s (Table 2). No statistically significant differences were observed between the cranberry extract and DMSO at any time (*p* > 0.05). The effect of exposure time (30 s versus 60 s) was not statistically significant for both solutions (*p* > 0.05 in all cases) in *S. oralis, A. naeslundii* and *V. parvula*.

For the secondary colonizer *F. nucleatum*, some effects on viable counts were observed after 30 s (*p* = 0.164) and after 60 s (decrease of 75.3%, *p* = 0.448), although not statistically significant. Additionally, no statistically significant reductions in viable counts were observed for DMSO (*p* > 0.05) (Table 2). No statistically significant differences were observed between the cranberry extract and DMSO at any time (*p* > 0.05). The effect of exposure time was however, statistically significant for the cranberry extract (*p* = 0.022) and DMSO (*p* = 0.035).

For the periodontal pathogens *A. actinomycetemcomitans* and *P. gingivalis*, no significant reductions in viable counts after 30 s or 60 s of exposure to the cranberry extract (*p* > 0.05) were observed when compared to negative control: reductions of 11.5% for *A. actinomycetemcomitans* and 39.3% for *P. gingivalis* after 60 s. The same was true for DMSO (*p* > 0.05) (Table 2). No statistically significant differences were observed in the effectiveness comparing the two solutions at applied times or when comparing exposure times (*p* > 0.05 for all cases).

#### 3.2.3. Antibacterial Effects in an in Vitro Biofilm Model: CLSM

The CLSM analysis showed that, after 72 h of incubation on HA surfaces, control biofilms covered the entire disc surface as a flat layer of cells combined with stacks of bacteria aggregations, showed a live/dead cell ratio (i.e., the area occupied by living cells divided by the area occupied by dead cells) of 1.43 (SD 0.10) and 1.25 (SD 0.15), after exposure of 30 and 60 s, respectively, to PBS (Figure 2a, b). Table 3 depicts the effects of the cranberry extract on the live/dead cell ratio of the whole biofilm obtained by CLSM. It could be observed that, after exposure of 30 s to cranberry extracts and to the 4% DMSO solution, cell vitality significantly decreased in the biofilms, showing, respectively, live/dead cell ratios of 0.67 (SD 0.07) and 0.77 (SD 0.04) for 4% DMSO (*p* < 0.001 in both cases, when compared to negative control biofilms) (Figure 2c, e). After 60 s of exposure (Figure 2f), reductions in viability were also statistically significant for cranberry extracts (live/dead cell ratio of 0.56 (SD 0.02), *p* < 0.001; Figure 2f) and for DMSO solution (live/dead cell ratio of 0.78 (SD 0.05), *p* < 0.001; Figure 2d), when compared to control biofilms (live/dead cell ratio of 1.25 (SD 0.15)). These results are consistent with those observed by means of qPCR, with significant differences in viable counts of initial and early colonizers, after exposure to cranberry extracts and DMSO solution, when compared to negative control biofilms. Statistically significant differences were observed between the cranberry extract and DMSO after 60 s of exposure (*p* = 0.027) (Table 3).



0.05, significant differences when comparing exposure times for an antimicrobial

**Figure 2.** Maximum projection of confocal laser scanning microscopy (CLSM) images of the whole biofilm, grown 72 h over hydroxyapatite surfaces, and stained with LIVE/DEAD® BacLightTM Bacteria Viability Kit, after exposure to: (**a**,**b**) negative controls, 30 and 60 s, respectively (phosphate buffer saline, PBS); (**c**,**d**) 4% dimethyl sulfoxide (DMSO) solution, 30 and 60 s, respectively; (**e**,**f**) cranberry extracts (20 g L<sup>−</sup>1), 30 and 60 s, respectively. (Scale bar = 100 μm).

**Table 3.** Effect of the cranberry extract on the live/dead cell ratio (i.e., the area occupied by living cells divided by the area occupied by dead cells) of the whole biofilm obtained by confocal laser scanning microscopy (CLSM). PBS: phosphate buffer saline; DMSO: 4% dimethyl sulfoxide solution.


Based on estimated marginal means; <sup>a</sup> *p* value, adjustment for multiple comparisons (Bonferroni).

#### *3.3. Anti-Biofilm Assay*

#### 3.3.1. Anti-Biofilm Assay: Bacteria Counts

The cranberry extract, at a concentration of 0.20 mg mL<sup>−</sup>1, significantly inhibited the incorporation of five of the six studied bacteria species in the in vitro biofilm model (Table 4). After 6 h of contact, and compared to negative control biofilms, two of the three initial and early colonizers were significantly reduced on the HA surfaces: 98.9% for *S. oralis* (*p* < 0.001) or 90.9% for *V. parvula* (*p* < 0.001), when exposed to cranberry extracts. No significant impact was observed for *A. naeslundii*.

Periodontal pathogens showed a similar trend. *P. gingivalis* showed the largest impact of cranberry extracts: 97.2% (*p* < 0.001), with counts of 1.1 <sup>×</sup> 10<sup>3</sup> (SD 1.1 <sup>×</sup> 103) CFUs mL−1, compared to 4.0 <sup>×</sup> 104 (SD 2.9 <sup>×</sup> 104) CFUs mL−1, in negative control biofilms. Reductions *A. actinomycetemcomitans* (84.0%) and *F. nucleatum* (75.4%) were statistically significant (*p* < 0.001 in both cases).

For DMSO, a significant impact was observed for the three periodontal pathogens and for *S. oralis*, when compared to control biofilms (*p* < 0.005 in all cases; Table 4).

Significant differences were observed in the effectiveness comparing the cranberry extract and DMSO solution after 6 h of biofilm evolution in *V. parvula* (*p* < 0.001) and *A. actinomycetemcomitans* (*p* = 0.024).


**Table 4.** Anti-biofilm effects of the cranberry extract on the mean number of bacteria counts, incorporated during the 6 h of devolvement in the in vitro multi-speciesbiofilm model (in colony forming units, CFUs mL−1, determined by quantitative real-time polymerase chain reaction (qPCR)). Data are expressed as mean

 and

#### 3.3.2. Anti-Biofilm Assay: CLSM

CLSM analysis showed that, after 6 h of incubation on HA surfaces, formed biofilms showed the typical features of bacteria communities in their first steps, with a high percentage of live cells versus dead cells, that was evidenced by a live/dead cell ratio of 1.44 (SD 0.01) (Figure 3a,b). The effect of the exposure of the biofilms for 6 h to the cranberry extract, at a concentration of 0.20 mg mL−1, was evident as it was not possible to observe well-structured biofilms, contrary what happened in the control samples. Although the biomass was reduced, no significant differences in bacteria vitality were observed when compared respect to controls (live/dead cell ratio of 0.99 (SD 0.01), *p* = 0.160; Table 3; Figure 3c,d), suggesting a limited antiseptic effect, and highlighting a desired effect on bacteria adhesion. Conversely, DMSO showed a similar live/dead cell ratio (1.047 (SD 0.14); Figure 3e,f), and biofilms were normally formed, with no significant differences when compared to control biofilms (*p* = 0.101; Table 3). These results are consistent with those observed by qPCR, which showed significant differences in bacteria counts of the tested bacteria species when biofilms formed in the presence of the cranberry extract where compared with control biofilms.

No statistically significant differences were observed in the effectiveness comparing the cranberry extract and DMSO at applied time (*p* = 1.000) (Table 3).

**Figure 3.** Maximum projection of confocal laser scanning microscopy (CLSM) images of the whole biofilm after 6 h of development, growing in the presence of 0.20 mg mL−<sup>1</sup> of cranberry extract, over hydroxyapatite surfaces, and stained with LIVE/DEAD® BacLightTM Bacteria Viability Kit, after exposure to: (**a**,**b**) negative control (phosphate buffer saline, PBS); (**c**,**d**) cranberry extract; (**e**,**f**) 4% dimethyl sulfoxide (DMSO) solution.

#### **4. Discussion**

Since bacteria resistance to antibiotics is becoming an increasing health threat worldwide, alternative strategies to prevent or limit biofilm formation are a relevant goal. A growing body of evidence has demonstrated that plant extracts offer relevant antimicrobial and anti-biofilm potentials, with no significant risk of increasing antibiotic resistances. A vast number of phytochemicals have been recognized as valuable alternatives and complementary products to manage bacterial infections [45,46]. Cranberry (*Vaccinium macrocarpum*) fruits are particularly rich in biologically active phenolic compounds and organic acids [13], as it has also been confirmed from our results (Table 1). Numerous in vivo and in vitro studies have showed that different cranberry compounds/fractions/extracts possess antibacterial properties (against both Gram-positive and Gram-negative bacteria species) on various pathogenic bacteria in urinary tract infections and other diseases [33,47–49]. In this context, the present study has confirmed the antibacterial capacity of cranberry extracts against the six bacteria species (*S. oralis* CECT 907T, *V. parvula* NCTC 11810, *A. naeslundii* ATCC 19039, *F. nucleatum* DMSZ 20482, *A. actinomycetemcomitans* DSMZ 8324, and *P. gingivalis* ATCC 33277) tested in planktonic state; this findings contradict, at least partially, those of La and co-workers [50], who concluded that A-type proanthocyanidins did not present any effect on *P. gingivalis* in planktonic state. In view of our results from the UPLC-DAD-ESI-TQ MS analysis of the cranberry extract, a possible explanation for this disagreement could be that the antiseptic effect comes, not only from these A-type proanthocyanidins, but also from some of the other components of cranberry extracts, such as phenolic acids. In this context, the necessity of carrying out previous compositional characterization of cranberry extracts should be noted—as we have done in our study—in view of the diversity in this kind of products that affects their bioactivity [51].

However, bacteria are normally arrange as biofilms, predominating these sessile communities in most of the environmental, industrial, and medical habitats [52]. In fact, these highly structured bacteria communities are found in the mouth, allowing bacteria cells to withstand the natural defence mechanisms, as well as the host's immune defences or the effects of antimicrobial agents [53–55]. Therefore, the study of the antimicrobial of cranberry extracts should be performed against bacteria organized in biofilms. The results of the present study indicate that, when testing bacteria organized in biofilms, bacteria viability was affected by exposure to the cranberry extract at 20 mg mL−<sup>1</sup> after 30 and 60 s of exposure. However, a significant effect was only observed for initial and early colonizers (*S. oralis, A. naeslundii*, and *V. parvula*), but, in agreement with other studies, not for periodontal pathogens (*F. nucleatum, P. gingivalis*, and *A. actinomycetemcomitans*). Philips and coworkers [56], in a recent investigation assessing the inhibitory effects of berry fruit extracts on *S. mutans* biofilms, indicated that bacteria viability was not significantly affected, as also concluded by Koo and coworkers [33]. Biofilms are an intriguing structure which demonstrate greater resistance to antimicrobial agents when compared to organisms in planktonic form [31]. A previous study using a cranberry juice concentrate formulated as a thermoreversible gel [11], showed antibacterial properties against *A. actinomycetemcomitans* and *P. gingivalis*, in contrast to the results of our study. The variability of the results may be due to the different types of samples and formulations used.

Besides the antibacterial effects, this investigation highlights new possible features regarding the anti-biofilm activity of cranberry extracts against periodontal pathogens. Bacteria adhesion to oral surfaces is the initial and crucial step in dental biofilm development and, therefore, in the pathogenesis of periodontal diseases. The cranberry extract, at a concentration of 0.20 mg mL−1, inhibited the colonization of the six tested bacteria species in the in vitro biofilm model, especially for periodontal pathogens *P. gingivalis* (97.2% of reduction), *A. actinomycetemcomitans* (84%), and *F. nucleatum* (75.4%), being the impact statistically significant (*p* < 0.001 in all cases), when compared to control biofilms. Additionally, initial and early colonizers were significantly affected: *S. oralis* (98.9%, *p* < 0.001) or *V. parvula* (90.9%, *p* < 0.001). Different studies have described the role of cranberry constituents in bacteria adhesion and biofilm development: Philips and coworkers [56] indicated that cranberry extracts were the most effective extract in disrupting *S. mutans* biofilm integrity and structural

architecture, without significantly affecting bacteria viability; La and co-workers [50] observed that A-type cranberry proanthocyanidins did not have any effect on *P. gingivalis* planktonic growth, but they did inhibit biofilm formation. The anti-biofilm effect of cranberry extracts in our biofilm model was also confirmed by CLSM, with a significant disturbance on biofilm structure, a qualitative assessment that was consistent with the quantitative data provided by qPCR.

Labreque and coworkers [37] and Yamanaka and coworkers [38] observed that the non-dialyzable constituent fraction of cranberry (NDM) interfered with the colonization of *P. gingivalis* and *F. nucleatum* in the gingival crevice, reducing bacteria coaggregation in periodontal diseases [37,38,57,58]. Moreover, Polak et al. [58] found that NDM adhesion of *P. gingivalis* and *F. nucleatum* onto epithelial cells, and NDM consumption by mice attenuated the severity of experimental periodontitis, compared with a mixed infection without NDM treatment. Furthermore, NDM increased the phagocytosis of *P. gingivalis*. In addition, cranberries were described to restrain the proteolytic activity of the red complex, specifically the gingipain activity of *P. gingivalis*, trypsin-like activity of *Tannerella forsythia*, and chemotrypsin-like activity of *Treponema denticola* [59]. Cranberry extracts have also demonstrated the inhibition of the productions some cytokines: Bodet et al. [59] or Polak et al. [58] observed that NDM eliminated TNF-a expression by macrophages that were exposed to *P. gingivalis* and *F. nucleatum*, without impairing their viability.

The hydrophobic character of the cranberry extract has made the experiments difficult, requiring the use of the organic solvent DMSO in the tests, in order to overcome such complications. However, some antimicrobial activity of DMSO at the selected concentration (4%) was observed, and therefore, it may have a possible contribution in the antibacterial activity of the extract under investigation. In this way, studies have tested different concentrations, ranging up to 10% [60–63]. However, when used as a solvent, there is no established criteria as to which is the most appropriate concentration, and the interpretation of its effects on the microorganisms with which it interacts is of great importance in view of its widespread use as solvent in therapeutic and pharmacological studies [60–63]. In the present study, the 4% DMSO concentration was selected as the one that ensured complete solubilisation of the cranberry extract with minimum antimicrobial effects. However, in any case, the results obtained in the present study make evident the need to standardize an appropriate concentration of DMSO, suitable for bacterial experiments, considering that there is a discrepancy in the findings of different studies on the antimicrobial effects of different concentrations of DMSO.

#### **5. Conclusions**

This study has demonstrated that the incorporation of bacteria into the biofilm was significantly interfered, including relevant periodontal pathogens, such as *P. gingivalis, A. actinomycetemcomitans* and *F. nucleatum*. Our results support the hypothesis that cranberry components may interfere in the phase of bacteria adherence, disabling or inhibiting the adherence of periodontal pathogens and, therefore, preventing bacterial colonization. This fact could interfere with biofilm formation and possibly helping to maintain homeostasis and, thus, to prevent periodontal diseases. Anti-biofilm activity of cranberry extracts in the present study could be attributed to the presence of polyphenols, specifically phenolic acids and A-type proanthocyanidins, which are known to inactivate glucosyl-transferase and fructosyl-transferase that catalyse the formation of glucan and fructan, respectively, which play prime roles in biofilm formation and maturation [31]. It has also been reported that the polyphenols in cranberries led to desorption of biofilm by interfering with bacteria coaggregation [64]. Moreover, cranberries are supposed to reduce periodontal-related symptoms by suppressing inflammatory cascades as an immunologic response to bacteria invasion.

Despite the limitations of the study, and the great effect caused by the DMSO solvent, the research performed has identified an important anti-biofilm effect of cranberry on periodontal bacteria and serve as a support for the development of further studies, assessing the most effective vehicle and the ideal concentration to be used, without causing adverse effects on oral tissues.

**Author Contributions:** M.C.S. and H.R.-V. contributed to conception and design of the study with the aid of E.F., B.B., M.V.M.-A., M.S., and D.H., analysis and interpretation of data and drafted the manuscript. E.F. performed the statistical analyses. E.F., B.B., M.V.M.-A., M.S., and D.H. critically revised the manuscript. All authors reviewed the original draft and read and approved the final manuscript.

**Funding:** This work was funded by the Spanish Ministry of Industry, Economy and Competitiveness (AGL2015–64522-C2-R project) and the Comunidad de Madrid (ALIBIRD-CM 2020 P2018/BAA-4343). It was also part of the activities of Dentaid Extraordinary Chair in Periodontal Research (*Cátedra Extraordinaria Dentaid en Investigación Periodontal*, University Complutense of Madrid, Spain).

**Acknowledgments:** We thank for their technical assistance to Luis M. Alonso and Alfonso Cortés, from the Centre of Microscopy and Cytometry from the Complutense University of Madrid.

**Conflicts of Interest:** The authors declare no conflict of interest. None of the funders had any role in designing and/or conducting of the study; collection, management, analysis and interpretation of the data; and preparation, review, or approval of the manuscript.

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Improvement of the Flavanol Profile and the Antioxidant Capacity of Chocolate Using a Phenolic Rich Cocoa Powder**

#### **Rocío González-Barrio \*, Vanesa Nuñez-Gomez, Elena Cienfuegos-Jovellanos, Francisco Javier García-Alonso and Mª Jesús Periago-Castón**

Department of Food Technology, Food Science and Nutrition, Faculty of Veterinary Sciences, Regional Campus of International Excellence "Campus Mare Nostrum", Biomedical Research Institute of Murcia (IMIB-Arrixaca-UMU), University Clinical Hospital "Virgen de la Arrixaca", University of Murcia, Espinardo, 30100 Murcia, Spain; vanesa.nunez@um.es (V.N.-G.); elena.cienfuegosjovellanos@gmail.com (E.C.-J.); fjgarcia@um.es (F.J.G.-A.); mjperi@um.es (M.J.P.-C.)

**\*** Correspondence: rgbarrio@um.es; Tel.: +34-868-88-96-41

Received: 12 January 2020; Accepted: 12 February 2020; Published: 14 February 2020

**Abstract:** Chocolate is made from cocoa, which is rich in (poly)phenols that have a high antioxidant capacity and are associated with the prevention of chronic diseases. In this study, a new production process was evaluated in order to obtain a dark chocolate enriched in (poly)phenols using a cocoa powder with an improved flavanol profile. The antioxidant capacity (Oxygen Radical Absorbance Capacity (ORAC) assay) and the flavanol profile (HPLC-DAD and HPLC-FL) was determined. The analysis of the enriched chocolate showed that the total flavan-3-ols (monomers) content was 4 mg/g representing a 3-fold higher than that quantified in the conventional one. Total levels of dimers (procyanidin B1 and B2) were 2.4-fold higher in the enriched chocolate than in the conventional, with a total content of 6 mg/g. The total flavanol content (flavan-3-ols and procyanidins) in the enriched chocolate was increased by 39% compared to the conventional one which led to a 56% increase in the antioxidant capacity. The new flavanol-enriched dark chocolate is expected to provide greater beneficial effect to consumers. Moreover, the amount of flavonols provided by a single dose (ca. 200 mg per 10 g) would allow the use of a health claim on cardiovascular function, a fact of interest for the cocoa industry.

**Keywords:** phenolic compounds; HPLC-DAD; fluorescence detection; flavan-3-ols; procyanidins; ORAC; (+)-catechin; (−)-epicatechin; dark chocolate

#### **1. Introduction**

In recent years, the consumption of cocoa and cocoa products has aroused greater interest due to their beneficial health effects. Cocoa beans and their derivate products are rich in (poly)phenols, which are associated with the prevention of diseases related to oxidative stress, such as cardiovascular diseases, carcinogenic processes, and neurodegenerative diseases [1–3]. However, the beneficial effects of cocoa (poly)phenols depend on the amount consumed, their bioavailability, and the biological activity of the conjugates formed [4]. Cocoa contains several classes of phenolic compounds among which, flavan-3-ols, procyanidins and anthocyanins [5]. Flavanols (flavan-3-ols and procyanidins) are the most studied compounds in cocoa and its derivatives for their beneficial health effects. The monomers (+)-catechin and (−)-epicatechin and the dimers procyanidin B1 and procyanidin B2 have been identified as the main flavanols in cocoa beans. Among the flavan-3-ols, the major one is (−)-epicatechin, and among the procyanidins, its dimer B2, ranging from 0.4 to 4.5 mg/g DW and 0.3 to 10 mg/g DW, respectively, depending on the geographic origin [6]. At much lower concentrations, (+)-catechin

and procyanidin B1 have also been identified in cocoa, ranging from 0.2 to 0.7 mg/g DW and 0.003 to 0.05 mg/g DW, respectively. In addition, trimers and even decamers of flavan-3-ols have also been identified in cocoa and dark chocolate [7,8].

The contents of (poly)phenols in cocoa beans depends on factors, such as variety (genotype) and origin, as well as post-harvest treatments [5,9,10]. The technological process related to the processing of cocoa beans for chocolate manufacturing affect the flavanol profile, both qualitative and quantitative [5,11]. Chocolate-making consists of a multistep process. The cocoa beans are cleaned and separated prior to industrial processing. Afterwards, they are roasted; a critical stage in the final product, since on the one hand guarantees the safety of the product and on the other hand the development of the cocoa flavors. But, it should also be noted that the high of roasting temperatures and time, lead to the loss of bioactive compounds of interest, such as (poly)phenols. Therefore, the optimization of the process to limit the loss of (poly)phenols can be of great importance for cocoa industry that wants to develop a healthy cocoa product, ensuring the safety and quality of the final product. Hence, several studies have been carried out to develop procedures to reduce the loss of phenolic compounds because there are many process that can significantly reduce the (poly)phenol content in cocoa [7,12,13]. It is known that the preservation of polyphenols during cocoa manufacture is important to the beneficial effects on human health associated with the consumption of cocoa and derived products. In a previous study, the process for obtaining a (poly)phenol enriched cocoa powder was described [14]. In this new process, some traditional stages, such as fermentation and roasting, were avoided and a step to inactivate the enzyme polyphenol oxidase was included, which helps preserve the polyphenol content in the raw cocoa bean.

The optimization of the industrial process can increase the percentage of (poly)phenols in the final products, such as cocoa powder, enabling them to be used as functional ingredients [15,16]. Several studies have shown that extracts from different cocoa beans and cocoa liquor have a powerful antioxidant activity, provided by the presence of flavonoids [8,17,18].

Many studies have described the benefits associated with the consumption of cocoa and its derivatives and this has increased the interest in obtaining functional cocoa products. As mentioned above, the benefits for human health depend not only on the amount of (poly)phenols consumed, but also on the bioavailability of the bioactive compounds [19]. In this regard, many studies have reported that the monomers, (+)-catechin and (−)-epicatechin, and the dimer procyanidin B2 are more bioavailable, being more beneficial than other flavanols with a greater molecular size [20]. The aim of this study was to evaluate the improvement in the flavanols profile and in the antioxidant capacity of the dark chocolate when using a (poly)phenol enriched cocoa powder, with the intention of yielding a functional product with potential beneficial effects for human health.

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

#### *2.1. Chemical Reagents and Standards*

All chemicals were purchased from Sigma Aldrich (St. Louis, MO, USA) while HPLC-grade solvents were purchased from either Scharlab (Barcelona, Spain) or Merck (Darmstadt, Germany). Standards of (+)-catechin and (−)-epicatechin (Sigma Aldrich, St. Louis, MO, USA) and procyanidin B1 and procyanidin B2 (Extrasynthese, Genay, France) were used for quantitative determinations.

#### *2.2. Development of (Poly)Phenol Enriched Dark Chocolate and Conventional Dark Chocolate*

Batches of 4 kg of (poly)phenol enriched chocolate and 4 kg of conventional dark chocolate were produced on a pilot scale (Figure 1). For the production of both chocolates, cocoa liquor (58%), sugar (20.5%), cocoa butter (6%), lecithin (0.45%), and vanillin (0.05%) were used, adding 15% (poly)phenol enriched cocoa powder for the enriched chocolate and 15% conventional cocoa powder for the conventional chocolate. The enriched cocoa powder was produced on an industrial scale from unfermented, blanched, and non-roasted cocoa beans (Amazonic-Trinitary variety-CCN51 clone from

Ecuador) using the procedure described by Cienfuegos-Jovellanos et al. [21]. Briefly, a blanching treatment with water (95 ◦C, 5 min) was applied to fresh cocoa beans, after removal of the pulp. This enriched cocoa cake was defatted by expeller pressing up to a fat content of 11%, and micronized to give a particle size of less than 75 microns. The conventional cocoa powder (Granada brand) was a commercial cocoa powder supplied by Natra Cacao S.L. (Valencia, Spain). It was produced from fermented and dried cocoa beans (Forastero variety from East Africa), being husked, roasted, defatted by hydraulic pressing up to a fat content of 11%, and micronized to a particle size of less than 75 microns. Both cocoa powders, the enriched and the conventional, had a moisture content of 5% and a pH of 5.6. The percentage of cocoa powder (15%) added was because in the sensory test no great differences were found between the enriched chocolate and the conventional one. This was confirmed by an internal tasting panel where a group of volunteers tasted different versions of chocolates with various percentages of enriched cocoa powder, and, apparently, the one with 15% added did not seem different from the standard chocolate and was acceptable in terms of overall sensory quality.

**Figure 1.** Diagram of the conventional and the new production process of (poly)phenol enriched cocoa powder [21] and chocolate.

The chocolate manufacturing process involved four steps: mixing/kneading, refining, conching, and tempering. In the first step, the ingredients (cocoa liquor, sugar, cocoa powder, and cocoa butter) were mixed together and kneaded, using refining or grinding procedures to provide a smooth chocolate paste. Then, the chocolate was subjected to the conching process, being mechanically kneaded to give it a more complete, homogeneous aroma and improved rheological characteristics. This step was carried out at temperatures between 75 ◦C and 80 ◦C, and other components—such as vanillin and lecithin—were added. Finally, the liquid chocolate was tempered, by cooling and heating under controlled conditions, and placed in molds.

#### *2.3. Determination of the Antioxidant Capacity*

The antioxidant capacity in cocoa powder and chocolate samples was determined by the Oxygen Radical Absorbance Capacity (ORAC) method, according to Cao et al. [22]. This method is based on the inhibition of the oxidation induced by a peroxy-radical, using a standard with antioxidant capacity as the substrate and a fluorescent probe to measure the signal. Fluorescein was used as the indicator, Trolox as the standard, and 2,2 -azobis(2-amidino-propane) dihydrochloride (AAPH) as the peroxyl radical generator. The assay was carried out using a fluorescent microplate reader (Synergy 2 Multi-Mode Microplate Reader, Biotek, Winooski, VT, USA) and 96-well black microplates equipped with a fluorescence filter having an excitation wavelength of 485 nm and an emission wavelength of 520 nm. For each calibration solution, the blank (0.075 M phosphate buffer, pH 7.0) and the samples were added to the corresponding wells. The plate reader also has an incubator and two injection pumps, which added the fluorescein and the AAPH during the assay; the temperature of the incubator was set to 37 ◦C. The fluorescence of each well was measured every 60 s for 90 min. The results were calculated using the standard curve provided by the instrument and expressed as μmol of Trolox Equivalents per g (μmol TE/g).

#### *2.4. Analysis of Flavanols by High Performance Liquid Chromatography with UV-Vis Detection (HPLC-DAD)*

The analysis of (+)-catechin and (−)-epicatechin monomers, as well as the dimers procyanidin B1 and procyanidin B2, in cocoa powder and chocolate samples was carried out by HPLC-DAD according to the method described by Andrés-Lacueva et al. [23], with some modifications. The chocolate samples were melted at 40 ◦C before defatting of the sample by the Soxhlet method, using petroleum ether at 55 ◦C for 24 h, and then drying of the sample at 40 ◦C. The defatted sample was homogenized in a vortex mixer with 5 mL of distilled water at 100 ◦C for 1 min. Twenty milliliters of methanol acidified with 0.1% HCl were added and agitated in a vortex for 2 min. Then, the homogenate was centrifuged at 1600*g* for 15 min, at 4 ◦C. This procedure was repeated twice, and the supernatants were combined and evaporated at 35 ◦C under vacuum to remove the methanol. The remaining aqueous extract was finally filtered through a 0.45-μm PTFE filter and analyzed by HPLC-DAD. The Agilent 1100 HPLC system was fitted with a quaternary pump, a degasser, a thermostatted column support, an autosampler, and a serial diode detector (Agilent Technologies, Waldbronn, GermanySpain). The separation of the different monomers and dimers was performed using a C18 Zorbax Eclipse XDB reverse phase column (150 × 2.1 mm, i.d. 5 μm) (Agilent Technologies, Spain) thermostatted at 35 ◦C. The mobile phases used were 0.1% aqueous formic acid (solvent A) and acetonitrile (B), at a flow rate of 0.6 mL/min. Elution began with a gradient from 4 to 10% B in 25 min, followed by a gradient to 13% B at 30 min, to 15% B at 33 min, and to 50% B at 35 min, followed by washing and then a return to the initial conditions at 45 min. Chromatograms were recorded at 280 nm. The identification of flavanols was performed by comparison of their retention times with those of the authentic standards of (+)-catechin hydrate (−)-epicatechin, procyanidin B1 and procyanidin B2. Quantification was based on calibration curves constructed using 5 to 25 ppm (+)-catechin, 20 to 100 ppm (−)-epicatechin, 1 to 10 ppm procyanidin B1, and 5 to 50 ppm procyanidin B2.

#### *2.5. Analysis of Flavan-3-ols and Procyanidins by High Performance Liquid Chromatography and Fluorescence Detection (HPLC-FL)*

The analysis of procyanidins in the cocoa powder and chocolate samples by HPLC-FL was carried out according to the method described by Gu et al. [8]. This included defatting of the sample (2 g) using 45 mL of n-hexane, followed by centrifugation and evaporation. After that, the extraction of the procyanidins was performed according to the methodology of Gu et al. [24]. Briefly, defatted samples were homogenized in a vortex mixer with acetone/water/acetic acid (70/29.5/0.5, *v*/*v*/*v*), sonicated at 37 ◦C for 15 min, and finally centrifuged at 3000 *g* for 5 min at 4 ◦C. The extraction process was performed twice more, each time adding 15 mL of the acetone/water/acetic acid (70/29.5/0.5, *v*/*v*/*v*) to the remnant. Finally, the supernatants obtained in each extraction stage were combined and the organic

phase was removed under vacuum. The aqueous phase remaining was shaken in an ultrasonic bath and purified using a vacuum-equipped solid phase extraction unit (SPE) (Merck, Darmstadt, Germany). A lipophilic filler of Sephadex LH-20 (Scharlab, Spain) was used as the solid phase. For each sample, 3 g of Sephadex LH-20 were packed into a column (6 × 1.5 cm) and preconditioned with 15 mL of methanol/water (30/70, *v*/*v*) overnight before use. The sample was passed through sorbent and collected. Once the procyanidins had been retained in the adsorbent, the column was washed with 40 mL of a mixture of methanol/water (30/70, *v*/*v*) to remove the sugars and other phenols. The procyanidins were then recovered with 80 mL of acetone/water (70/30, *v*/*v*) until the cartridge was completely clean. Finally, the acetone was evaporated at 45 ◦C under vacuum. The aqueous extract obtained was lyophilized and dissolved in 10 mL of acetone/water/acetic acid (70/29.5/0.5, *v*/*v*/*v*), before being filtered and injected into the HPLC-FL system. The different procyanidin fractions were separated, according to their degree of polymerization, in the same Agilent 1100 HPLC system described above. Separation of the different fractions was achieved on a C-8 normal phase column, Luna Silica 100 Å (150 × 4.6 mm, i.d. 5 μm) (Phenomenex, Madrid, Spain). The mobile phase used was a mixture of dichloromethane (A), methanol (B), and glacial acetic acid/water (50/50, *v*/*v*) (C) at a flow rate of 0.5 mL/min. Elution began with a linear gradient from 14 to 28% B in 30 min, followed by a linear gradient from 28 to 39% B at 45 min, and from 39 to 86% B at 50 min. From 50 to 55 min an isocratic gradient of 86% B was used, followed by washing and then a return to the initial conditions at 70 min. A presence of 4% C remained constant throughout the elution. A fluorescence detector, set at an excitation wavelength of 276 nm and an emission wavelength of 316 nm, was used for detection. Due to the lack of commercially available standards, the quantification of the different procyanidin fractions was performed using previously published relative response factors for each one [25]. Briefly, the response factor of (−)-epicatechin was measured based on a calibration curve constructed using a commercial standard. Then, the response factors of the different procyanidin fractions were estimated using the corresponding relative response factors (0.65 for dimers, 0.69 for trimers, 0.61 for tetramers, 0.58 for pentamers, 0.45 for hexamers, 0.62 for heptamers, 0.52 for octamers, 0.36 for nonamers, 0.56 for decamers, and 0.45 for >decamers), which were used to calculate the concentration (mg/g) of each fraction.

#### *2.6. Statistical Analysis*

All analytical parameters were determined in triplicate for each sample, except the procyanidins analysis by HPLC-FL which was determined in duplicated. The data were expressed as means ± standard deviations (SD). Independent-samples T-tests were applied to determine the differences between means for all analyzed parameters (*p* < 0.05). The statistical analysis was carried out using GraphPad Prism version 6.02 for Windows, GraphPad Software (La Jolla, CA, USA).

#### **3. Results**

#### *3.1. Antioxidant Capacity*

The ORAC method gave values of 686 μmol TE/g for the conventional cocoa powder and 2861 μmol TE/g (more than 4-times higher) for the (poly)phenol enriched cocoa powder. A value of 412 μmol TE/g was obtained for the conventional dark chocolate and a value of 641 μmol TE/g (1.6-times higher) was obtained for the (poly)phenol enriched chocolate (Figure 2).

#### *3.2. Flavanol Analysis by HPLC-DAD*

The qualitative and quantitative profiles of cocoa powder and dark chocolate (enriched and conventional) were analyzed by HPLC-DAD. The chromatogram obtained for the dark (poly)phenols enriched chocolate (Figure 3) provides evidence that in all cases the separation and quantification of (−)-epicatechin and (+)-catechin monomers and procyanidin B2 and B1 polymers were achieved.

**Figure 2.** Oxygen Radical Absorbance Capacity (ORAC) antioxidant capacity of conventional and enriched cocoa powder (**a**) and chocolate (**b**). \* Indicates significant differences at *p* < 0.05.

**Figure 3.** HPLC-DAD chromatogram at 280 nm of dark (poly)phenol enriched chocolate. 1: Procyanidin B1; 2: (+)-catechin; 3: procyanidin B2; 4: (−)-epicatechin.

The results obtained for the cocoa powder and chocolate (conventional and enriched) are shown in Table 1. In both samples (cocoa powder and chocolate) the main flavanols found were (−)-epicatechin and procyanidin B2.


**Table 1.** Flavanol contents (mg/g) in the conventional and enriched dark chocolate analyzed by reverse phase HPLC-DAD 1.

<sup>1</sup> The values are the average of three replicates (*<sup>n</sup>* <sup>=</sup> 3) ± SD. \*, \*\*, \*\*\* Indicates significant differences at *<sup>p</sup>* <sup>&</sup>lt; 0.05, *<sup>p</sup>* <sup>&</sup>lt; 0.01 and *p* < 0.001, respectively, between conventional and enriched for each products (cocoa powder and chocolate).

Total monomers content was 12-fold higher in the enriched cocoa powder than that quantified in the conventional one, while the total dimers content was 15-fold higher. In the enriched chocolate, the total monomers content was 3-fold higher than in the conventional, while the total dimers content was 2.4-fold higher. Moreover, the total content of monomers and dimers was 3.9 mg/g for the conventional chocolate and 10.0 mg/g (2.6-times greater) for the enriched chocolate (Table 1).

#### *3.3. Procyanidins Analysis by HPLC-FL*

Due to the absence of standards for some of the different fractions of procyanidins, their quantification was performed using relative response factors previously published for chocolate [25]. Figure 4 show the chromatograms obtained for the (poly)phenol enriched chocolate. The method used allowed the separation of the different fractions, from monomers to decamers; the procyanidins of a higher degree of polymerization (>10) were quantified at the end of the chromatogram as a single peak.

**Figure 4.** HPLC-FL chromatogram of (poly)phenol enriched cocoa (poly)phenol enriched chocolate (b) 1: monomers; 2: dimers; 3: trimers; 4: tetramers; 5: pentamers; 6: hexamers; 7: heptamers; 8: octamers; 9: nonamers; 10: decamers; >10: procyanidins with a degree of polymerization of more than 10 units.

Table 2 shows the flavanol contents, covering the range from monomers to polymers, in the conventional and (poly)phenol enriched cocoa powder and in the conventional and enriched chocolate. In both the (poly)phenol enriched cocoa powder and enriched chocolate, the monomers were the main fraction, representing 23% of the total flavanols. However, in the conventional cocoa powder and conventional chocolate, the main fraction was comprised of the procyanidins with a degree of polymerization of more than 10 units (>10), this fraction representing 37% and 22%, respectively, of the total flavanols. Overall, the content of procyanidins in the enriched cocoa powder was 4.7-fold greater than in the conventional cocoa powder, and in the enriched chocolate it was 1.4-fold greater than in the conventional chocolate.


**Table 2.** Flavanol contents (mg/g) of the conventional and enriched cocoa powder and chocolate analyzed by normal phase HPLC-FL a.

<sup>a</sup> The values are the average of two replicates (*<sup>n</sup>* <sup>=</sup> 2) ± SD. nd (not detected). \*, \*\*, \*\*\* Indicates significant differences at *p* < 0.05, *p* < 0.01, and *p* < 0.001, respectively, between conventional and enriched for each product (cocoa powder and chocolate).

#### **4. Discussion**

A dark chocolate rich in (poly)phenols has been produced using an enriched cocoa powder, whose new production process reduces the losses of bioactive compounds, allowing these compounds to be preserved during chocolate production. The difference between the enriched and conventional cocoa powders lies in the processing, since the enriched cocoa powder was obtained from a new process in which some stages of the traditional process were replaced, such as fermentation, grain roasting, and defatting by hydraulic pressing. Conversely, other common processes used in the food industry were incorporated to minimize the loss of phenolic compounds, such as scalding to inactivate the enzyme (poly)phenol oxidase (PPO), deffating by expellers at low and controlled temperatures, and heat treatment by steam currents in an autoclave (Figure 1). The dose of cocoa powder added (15%) for the manufacturing of the chocolate samples was chosen on the grounds of a sensory test, in order to obtain no perceivable differences between the enriched chocolate and the conventional one. In spite of the astringency and the purple color of the enriched cocoa powder, this dose did not change the organoleptic properties of the enriched chocolate. Moreover, the enriched cocoa powder had a very plain profile, with no cocoa taste or aroma, since it had not undergone the roasting process; therefore, the precursors of the cocoa aroma and flavor were not developed since it was not fermented either.

The antioxidant capacity of the enriched cocoa powder used for the production of the dark chocolate rich in (poly)phenols was 2861 μmol TE/g representing 317% higher than that of the conventional cocoa powder. This shows the enhanced value of this (poly)phenol enriched cocoa powder. The antioxidant capacity of the enriched chocolate was 641 μmol TE/g representing 56% higher than that of the conventional chocolate, but both values are significantly higher than the corresponding values found in various fruit powders—acai (400 μmol TE/g), blueberry (260 μmol TE/g), cranberry (310 μmol TE/g), and pomegranate (190 μmol TE/g)—and in natural, non-alkalized cocoa powder (634 μmol TE/g) [17].

Taking into account that the chocolates were elaborated with 15% cocoa powder (enriched or conventional), theoretically, the antioxidant capacity should be 103 μmol TE/g for the conventional chocolate and 429 μmol TE/g for the enriched one. However, the values obtained were 4-fold higher for the conventional chocolate (412 μmol TE/g) and 1.5-fold higher for the enriched one (641 μmol TE/g). The differences observed between the hypothetical and the analytical results could be explained, in part, by the presence of other bioactive compounds in the cocoa liquor added (58%) during the chocolate manufacturing [5], which could increase the antioxidant capacity in the chocolate. In addition, the results obtained demonstrate that the enriched powder is less stable than the conventional one during chocolate processing. However, the antioxidant capacity of the enriched chocolate formulation, including 15% (poly)phenols enriched cocoa powder, was increased by 56% compared to the chocolate formulated with 15% conventional cocoa powder.

To achieve beneficial effects on human health, the daily recommended intake of antioxidants is 3000–3600 μmol TE [26]. However, the recommended five pieces of fruit and vegetables per day do not reach 50% of this value, providing between 1200 and 1640 μmol TE [26]. For this reason, the consumption of functional products enriched in (poly)phenols could be a strategy to increase the daily intake of antioxidants. Specifically, a 10-g portion of the enriched dark chocolate described here will provide a high level of antioxidants, more than 6000 μmol TE.

The analysis of the enriched chocolate measured by HPLC-DAD, showed that the total flavan-3-ols content was 4 mg/g representing a 186% higher than that of the conventional chocolate. Moreover, the content in enriched chocolate was also higher than that determined by other authors in dark chocolate with values ranging to 0.24–0.45 mg/g [27]. The results showing that, for chocolate, (−)-epicatechin was the second most abundant of the flavanols analyzed agree with other work reporting that one of the main compounds in cocoa beans is (−)-epicatechin [28]. In this study, the sum of the monomers, (−)-epicatechin and (+)-catechin, in the enriched chocolate (4.0 mg/g) was higher compared to the conventional one (1.4 mg/g), and was higher than that obtained by other authors for a dark chocolate (1.9 mg/g) [8]. These results show that, in the new functional product, there is an increase in the more bioavailable flavanols and therefore in the health benefits of the dark (poly)phenol enriched chocolate, compared to the conventional chocolate. This is because (+) catechin, (−)-epicatechin and procyanidin B2, which represent 43% of the total procyanidins, are more bioavailable than the other compounds, as some authors have shown, since their absorption in the gastrointestinal tract is higher [20,29–32].

The total flavanol content, covering the range from monomers to polymers, of the enriched cocoa powder was 78 mg/g representing 364% higher than that obtained for the conventional cocoa powder in this study and 90% higher than that obtained for a conventional cocoa powder by other authors who reported a value of 41 mg/g [8]. The mean contents quantified for the flavan-3-ols and procyanidins in this study are similar to the values published for a cocoa powder that was not affected by post-harvest variables [12,33]. Similarly, it should be noted that the total procyanidins content quantified in the (poly)phenol enriched cocoa powder was 1.3-fold higher than that quantified by Kealey et al. [12] in a cocoa powder obtained from fresh unfermented beans, lyophilized and defatted by Soxhlet, and 2.8-fold greater than that quantified by Misnawi et al. [33] in a cocoa powder obtained from fresh unfermented beans partially defatted using an expeller.

Gu et al. [8], using a similar method of extraction, purification with Sephadex LH20, and normal phase analysis using HPLC-ESI/MS with a fluorescence detector, reported values of total procyanidins from 8.5 to 20 mg/g in conventional dark chocolate samples - very similar to the 15 mg/g for conventional chocolate and the 21 mg/g in enriched chocolate found in this study. In addition, the total procyanidin value of 21 mg/g in cocoa powder samples recorded by Miller et al. [34] is close to that found here for conventional cocoa powder (17 mg/g). However, the (poly)phenol enriched chocolate developed in the present study has also been characterized by values of low-molecular-weight procyanidins higher than previously published data ranging values to 1.14–1.77 mg/g [8,35]. The sum of the classes ranging from monomers to hexamers found in the enriched chocolate was 17 mg/g, compared to 10 mg/g found by Gu et al. [8] in a dark chocolate and 11 mg/g found in this work for conventional chocolate. Theoretically, as both chocolates (enriched and conventional) were elaborated with 15% cocoa powder, the total flavanol content should be 11.7 mg/g for the enriched chocolate and 2.5 mg/g for the conventional, 2-fold and 12-fold higher than that quantified for the enriched chocolate and for the conventional one, respectively. These results could be explained, in part, by the presence of flavanols in the cocoa liquor added (58%) during the chocolate manufacturing.

#### **5. Conclusions**

Our results describing the formulation of a new dark chocolate enriched with (−)-epicatechin and procyanidin B2 was successfully achieved, as well as a notable enrichment of the oligomeric procyanidin fractions (from monomers to hexamers), when compared with a conventional dark chocolate. These results are of current interest to both large food companies and health professionals. The manufacturing of the enriched dark chocolate following the procedures described above would allow the use of a health claim related to cocoa flavanols. According to the Commission Regulation (EU) 2015/539 of 31 March 2015 [36], the new enriched chocolate could have the claim '*Cocoa flavanols help maintain the elasticity of blood vessels, which contributes to normal blood flow*'. In order to obtain this beneficial effect, 10 g of the new enriched dark chocolate should be consumed daily, which would provide more than 200 mg of total flavanols (flavan-3-ols and procyanidins ranging from dimers to decamers). In addition, the new dark chocolate formulation, including 15% (poly)phenols enriched cocoa powder, increased greatly the antioxidant properties compared to a conventional dark chocolate, while maintaining the organoleptic properties unchanged. However, a future consumer preference study could be of great interest to determine the acceptability of the new functional chocolate.

**Author Contributions:** Conceptualization, R.G.-B., E.C.-J. and M.J.P.-C.; Formal analysis, V.N.-G. and E.C.-J.; Methodology, R.G.-B. and E.C.-J.; Writing—original draft, R.G.-B. and V.N.-G.; Writing—review & editing, R.G.-B., E.C.-J., F.J.G.-A. and M.J.P.-C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank F.A. Tomás-Barberán from CEBAS-CSIC (Murcia, Spain) for his technical advice regarding the analysis of flavan-3-ols and procyanidins. All cocoa powder and chocolate samples were kindly provided by Natraceutical Group, Valencia, Spain.

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

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).
