Taste perception of cyclic oligosaccharides: α, β, and γ cyclodextrins

Abstract

Oligosaccharides, a subclass of complex carbohydrates, occur both naturally in foods and as a result of oral starch digestion. We have previously shown that humans can taste maltooligosaccharides (MOS) and that their detection is independent of the canonical sweet taste receptor. While MOSs most commonly occur in a linear form, they can also exist in cyclic structures, referred to as cyclodextrins (CD). The aim of this study was to investigate how the structure of the MOS backbone (i.e. cyclic form) and the size (i.e. degree of polymerization; DP) affect their taste perception. We tested taste detection of cyclodextrins with DP of 6, 7, and 8 (i.e. α-, β-, and γ-CD, respectively) in the presence and absence of lactisole, a sweet receptor antagonist. We found that subjects could detect the taste of cyclodextrins in aqueous solutions at a significant level (P < 0.05), but were not able to detect them in the presence of lactisole (P > 0.05). These findings suggest that the cyclodextrins, unlike their linear analogs, are ligands of the human sweet taste receptor, hT1R2/hT1R3. Study findings are discussed in terms of how chemical structures may contribute to tastes of saccharides.

Introduction

Complex carbohydrates are abundant in the human diet, where they serve as sources of energy, as prebiotics, and as dietary fibers. However, their detection by the gustatory system has remained historically overlooked, likely because saccharides larger than simple sugars (i.e. mono- and di-saccharides) were assumed to be tasteless to humans (Ramirez 1991). More recently, studies in our lab have shown that humans can taste short-chain, linear starch hydrolysis products, known as maltooligosaccharides (MOS) (Lapis et al. 2014, Lapis et al. 2016; Pullicin et al. 2017). We have also reported that the taste perception of MOS with a degree of polymerization (DP) greater than 3 is independent of the canonical sweet taste receptor hT1R2/hT1R3 (Lapis et al. 2016; Pullicin et al. 2017) and that these stimuli are described as “starchy” (Lapis et al. 2016; Pullicin et al. 2017). These findings support earlier work reporting that rodents can taste MOS and their larger counterparts, maltopolysaccharides (MPS) (Nissenbaum and Sclafani 1987; Sclafani et al. 1987; Sclafani 1988; 1991; Giza et al. 1991; Treesukosol et al. 2009; Spector and Schier 2018). Though there has been some work investigating the gustatory mechanisms underlying the taste detection of MOS and/or MPS in humans (Lapis et al. 2016, Pullicin et al. 2017) and rodents (Sclafani et al. 2007; Zukerman et al. 2009; 2013; Spector and Schier 2018), the exact mechanism has not yet been defined.

MOSs are composed of glucose subunits connected via α-1–4 glycosidic bonds with a DP of 3–20 (Astray et al. 2009) (see Fig. 1A). They can exist in linear form, which we have produced (Balto et al. 2016; Pullicin et al 2017; Pullicin et al. 2018) and used in our previous studies (Balto et al. 2016; Lapis et al. 2016; Pullicin et al. 2017; Pullicin et al. 2018). They are also found in branched form, in which linear structures are connected to each other via α-1–6 glycosidic bond (see Fig. 1B); the effect of branching on taste detection of oligosaccharides is currently unknown. The prevalence of linear vs. branched oligosaccharides produced from starch depends on both the starting material (e.g. linear amylose vs. branched amylopectin) and the method of hydrolysis (Lee et al. 2013; Zhao et al. 2017; Gangoiti et al. 2020).

Fig. 1.

Molecular structures of maltooligosaccharides (MOSs). Chemical structures of (A) linear MOS, (B) branched MOS, and (C) cyclodextrins.

While not as common, MOS also occur in cyclic structures with 6, 7, or 8 glucose subunits, which are referred to as α-, β-, and γ-cyclodextrins, respectively (Szejtli 1998) (see Fig. 1C). Cyclodextrins are produced commercially for their applications in food, pharmaceutical, and consumer products (Biwer et al. 2002; Farahat 2020), although they can be found rarely in nature. Cyclodextrins are toroidal, i.e., shaped in a cone with one larger and one smaller opening. The molecule’s hydroxyl appendages are outside its cavity, thus giving the molecule a relatively hydrophilic exterior and relatively hydrophobic interior (Kurkov and Loftsson 2013). Due to this unique structure, they form inclusion complexes with hydrophobic compounds (Linde et al. 2009). Consequently, cyclodextrins are typically used to increase the solubility and/or stability of the guest compound with significant hydrophobic character (e.g. active ingredient) and/or to reduce its unpleasant taste or smell (Biwer et al. 2002; Marcolino et al. 2011; Morrison et al. 2013). The guest/target compounds are released upon degradation of the cyclodextrin in the gastrointestinal tract. Research suggests that β-cyclodextrin is often the most effective at forming inclusion complexes and is the easiest to purify due to its relatively low solubility in water (Marcolino et al. 2011; Farahat 2020).

Cyclodextrins differ from linear MOSs in a few ways. Most notably, their structure is conical, whereas linear MOSs tend to form a relaxed helix in solution with DP 6 making a full helical turn (Hayashi et al. 1981) (see Fig. 2). Another important structural difference between cyclodextrins and MOS is that the former is a nonreducing sugar while the latter is a reducing sugar. Finally, while both cyclodextrins and MOSs go through a stepwise enzymatic hydrolysis process into glucose after oral administration (Dona et al. 2010), cyclodextrins are more resistant to enzymatic hydrolysis than their linear counterparts due to their cyclic structure (Kurkov and Loftsson 2013). Specifically, α-amylase, an endo-acting enzyme that catalyzes the hydrolysis of α-1,4 glycosidic linkages, comprises 5–7 binding subsites and a catalytic site (see Lim and Pullicin 2019). Because cyclodextrins form cones, their glycosyl residues curve away from the binding subsites of α-amylase, and it is therefore less effective at cleaving them (Jodái et al. 1984; Jodál et al. 1984; Lumholdt et al. 2012).

Fig. 2.

Shapes taken by (A) linear maltooligosaccharide and (B) cyclodextrins in solution.

Cyclodextrins have been generally described as “slightly sweet” (Szejtli and Szente 2005; Zakharova et al. 2016), though no formal data support this claim; it is possible that previous reports of sweetness could be due to their partial enzymatic hydrolysis into disaccharides via salivary α-amylase. The objective of this study was to investigate how the structure of the MOS backbone (i.e. cyclic form) and the size (i.e. DP) affect their taste perception. To do this, we initially tested the taste detection of α-, β-, and γ-cyclodextrins in the presence and absence of lactisole, a known sweet–taste blocker (Jiang et al. 2005).

Experiment 1

Materials and methods

Subjects

A total of 27 subjects (18 females, 9 males) between 20 and 58 years of age (mean = 30 years old) were recruited from the Oregon State University campus. Subject inclusion criteria were individuals who (i) are generally healthy; (ii) are nonsmoker; (iii) are not pregnant; (iv) do not take any prescription medication including insulin; (v) have no history of taste or smell loss, or other oral disorders (e.g. burning mouth syndrome); (vi) have no current oral lesions or piercings; (vii) are free from food-related allergies. Respondents who met all of the above criteria were invited to participate in the study. Prior to the test session, subjects were also asked to comply to the following restrictions: (i) no dental work within 48 h; (ii) no alcohol consumption within 12 h; (iii) no consumption of foods and beverages that were acidic or caffeinated and/or contain dairy within 4hrs; (iv) no consumption of food or beverage of any kind except water within 1 h; and (v) no use of any menthol-containing products (e.g. toothpaste, mouthwash, and chewing gums) within 1 h. The Oregon State University Institutional Review Board approved the experimental protocol. Subjects gave written informed consent and were compensated for their participation.

Stimuli

A total of 4 test stimuli were used for the experiment: (i) α-cyclodextrin (DP 6), (ii) β-cyclodextrin (DP 7), (iii) γ-cyclodextrin (DP 8), and (iv) glucose (DP 1) as a sweet tasting standard. The cyclodextrins were provided by Cyclolab (Budapest, Hungary), and glucose was purchased from Spectrum Chemical (Gardena, CA). Lactisole was purchased from Domino Specialty Ingredients (Cypha lactisole; Yonkers, NY), and acarbose was purchased from MuseChem (Fairfield, NJ). All test stimuli were food grade and presented at 75 mM. Based on our previous studies (Pullicin et al. 2017), glucose and MOS samples at 75 mM elicit a weak but recognizable taste sensation for most subjects. Importantly, we noticed that 75 mM β-cyclodextrin never fully dissolves. Accordingly, the supernatant was used as a taste solution. Previous work suggests that the solubility of β-cyclodextrin in water is approximately16mM (Chatjigakis et al. 1992), and systematic testing in our lab confirmed that β-cyclodextrin is saturated at 18.5 mM in room-temperature water. Thus, the β-cyclodextrin solution used for sensory testing can be assumed to be about 18.5 mM.

The test stimuli were prepared with and without 1.4 mM lactisole. Lactisole is an antagonist for the canonical sweet receptor, as it binds to the transmembrane domain of T1R3 and prevents further activation by sweet stimuli (Jiang et al. 2005). While cyclodextrins are more resistant to the enzymatic hydrolysis by salivary α-amylase (Kondo et al. 1990), 5 mM acarbose was added to all stimuli and corresponding water blanks to prevent any potential hydrolysis. It was found that 5 mM acarbose was sufficient in inhibiting α-amylase previously (Lapis et al. 2016). Test stimuli were prepared weekly with deionized (DI) water and stored in airtight glass containers at 4–6°C. Stimuli were prepared at least 16 h prior to the test session to allow for complete mutarotation of glucose tautomers (Pangborn and Gee 1961).

Procedure

Each subject participated in a single session. At the beginning of the session, subjects were verbally instructed on the task that they were to perform.

Taste discrimination task.

A triangle test method was used, where 1 target stimulus and 2 blanks were given for each trial. Before testing, subjects were asked to rinse their mouth with DI water and spit into a sink 3 times. Subjects were then asked to extend the tongue out of the mouth and hold it immobile between the lips. In a random, sequential manner, a set of 3 samples were applied by the experimenter rolling saturated cotton swabs across the tip of the tongue (volume of stimulus absorbed by swab is ~0.15 mL). Subjects were instructed to not put the tongue back inside the mouth to avoid touching the roof of mouth while tasting the sample. Once they had tasted the sample, they then put their tongue back into the mouth and rinsed immediately with the rinse water before the next sample. After the 3 samples were tasted, subjects indicated, on a provided paper ballot, which was the odd sample. To prevent any potential olfactory cues, subjects wore nose clips while performing the task. Although lactisole itself is tasteless, it has been known to elicit a sweet “water-taste” sensation when rinsed away from the receptor (Galindo-Cuspinera et al. 2006); data suggest this “water-taste” effect is attenuated in colder temperatures (Green and Nachtigal 2015; Lapis et al. 2016). Thus, all target stimuli, blanks, and rinse water were presented at 10°C to prevent sweet “water-taste” (Green and Nachtigal 2015).

Practice/screening trials.

Two practice/screening trials, one with 150 mM glucose and another with 75 mM maltose were given. To advance to the test trials, subjects were required to successfully discriminate 150 mM glucose from blanks. Two participants did not pass the screening trial.

Test trials.

During the test trials, the test stimuli (α-, β-, γ-cyclodextrins, and glucose) were given to subjects in the same manner as in the practice trials. The stimuli were presented in 2 test blocks: (i) without and (ii) with lactisole. Presentation order of the test blocks, the 4 stimuli within the block, and 3 stimuli within a triangle test were randomized and counter-balanced across subjects. For testing, all target stimuli, blanks, and rinse water were served at 10 °C to attenuate the sweet “water-after taste” effect of lactisole. Subjects were given a 1-min break in between each set of 3 samples during which they rinsed their mouths with cold water. After subjects completed 4 triangle tests within one block, they were given a 3-minute break, during which they were asked to rinse their mouths 3 times with cold water. The experiment was conducted on a one-on-one basis in the psychophysical testing laboratory in the Department of Food Science and Technology at Oregon State University.

Data analysis

Proportion correct was calculated for each stimulus at each condition (i.e. with or without lactisole). These proportions were converted to corresponding d’ values for a triangle test (Ennis 1993). Briefly, the d’ value represents the degree of perceptible difference between 2 stimuli, and is a measure of the separation of the signal (target stimulus) and noise (blank) distributions in terms of the standard deviations of the distributions. The d’ analysis was used to determine significant discrimination at P-value of less than 0.05.

Results

Figure 3 shows the d’ (discriminability) values for glucose and the cyclodextrins tested. We found that subjects could detect all 4 test stimuli at a similar degree (proportion correct = 0.56–0.68, d’ = 1.77–2.39) in the absence of lactisole. However, when lactisole was added, the taste detection decreased significantly for all stimuli tested (P < 0.05), such that all stimuli were not detectable at a significant degree (proportion correct = 0.28–0.40, d’ = 0–0.88, P > 0.1). This finding suggests that, unlike linear MOSs tested in our previous studies (Lapis et al. 2016; Pullicin et al. 2017), the taste detection of cyclodextrins is dependent on the hT1R2/hT1R3 sweet receptor. It is worth noticing that the detection of γ-cyclodextrin in the presence of lactisole remained slightly higher (proportion correct = 0.40, d’ = 0.88, p = 0.1) than α-, β-cyclodextrin, or glucose (proportion correct = 0.28–0.32, d’ = 0). We hypothesized that γ-cyclodextrin is a sweet material, but its sweet taste was incompletely blocked by lactisole. The latter hypothesis is supported by a previous report showing that γ-cyclodextrin effectively form complexes with gymnemic acid, another known sweet taste blocker (Izutani et al. 2005).

Fig. 3.

Discriminability of glucose, α-, β-, and γ-cyclodextrin at 75 mM, in the presence and absence of the hT1R2/hT1R3 inhibitor lactisole. Stimuli at 10°C were swabbed on the tip of the subject’s tongue. Subjects performed triangle tests to identify the target sample from 2 blanks. The gray horizontal line indicates significant discriminability (P < 0.05). Asterisks indicate a significant difference in discriminability between the 2 conditions (P < 0.05).

Experiment 2

In this follow-up experiment, we tested the hypothesis that γ-cyclodextrin is a sweet taste stimulus and hence its taste is indiscriminable from that of glucose.

Materials and methods

Subjects

Twenty-four subjects (11 females, 13 males) aged between 18 and 57 years of age (mean = 26) were recruited from the Oregon State University campus. Ten of those also participated in Experiment 1. The inclusion criteria were the same as in the previous experiment, and subjects were asked to follow the same restrictions before testing.

Stimuli

Glucose and γ-cyclodextrin were used for this experiment. These stimuli were prepared at 2 different concentrations each: a low concentration (75 mM γ-cyclodextrin, 150 mM glucose) and high concentration (120 mM γ-cyclodextrin, 280 mM glucose). Seventy-five millimolar γ-cyclodextrin was selected for the low concentration since it was used in Experiment 1. One hundred and twenty millimolar γ-cyclodextrin was used for the high concentration. Iso-intense concentrations of glucose were determined through a series of same/different judgment between a target γ-cyclodextrin (75 mM; 120 mM) and 1 of 4 glucose concentrations within each concentration level (100, 120, 150, 170 mM; 260, 280, 300, 320 mM). Five individuals participated and the concentration pairs that received the most “same” responses were chosen as iso-intense concentrations. As in Experiment 1, 5 mM acarbose was added to all stimuli to prevent potential hydrolysis by salivary α-amylase. Test stimuli were prepared weekly with DI water and stored in airtight glass containers at 4–6°C, and all target stimuli, blanks, and rinse water were presented at 10°C. Stimuli were prepared at least 16 h prior to the test session to allow for complete mutarotation of glucose tautomers (Pangborn and Gee 1961).

Procedure

Each subject participated in a single session, which was structured in the same way as in Experiment 1. To advance to the test session, subjects were required to successfully discriminate 150 mM glucose. All participants advanced to the discrimination task. During the test session, 2 trials of triangle test were given.

Taste discrimination task.

Within each trial, a set of 3 samples (1 γ-cyclodextrin and 2 glucose samples) were applied in sequence by rolling saturated cotton swabs across the tip of the tongue. If subjects correctly discriminated either set, then they were asked to discriminate the same set, with a different order, a second time. If they correctly discriminated the sample once again, they were asked what the discriminable feature(s) of the sample was/were. This was done to ensure that successful discrimination was not due to intensity differences. Only the first 2 discrimination tests (1 at low concentration, 1 at high) were considered data, and subsequent testing is reported but has not been included in the analysis.

Note that in Experiment 2, subjects were asked to roll stimuli on their own tongues, instead of having the experimenter apply the stimuli. This was done to comply with special measures for COVID-19 safety protocols. This restriction was not applicable for Experiment 1, which was conducted under different COVID-19 guidelines.

Data analysis

Data were analyzed using d’ to determine whether the subjects were able to significantly discriminate the test stimuli. Proportion correct was calculated for each stimulus at each condition. These proportions were converted to corresponding d’ values for a triangle test (Ennis 1993), and d’ analysis was used to determine significant discrimination at P-value of less than 0.05.

Results

Figure 4 shows the d’ (discriminability) values for glucose vs. γ-cyclodextrin at low and high concentrations. We found that, at both concentrations, subjects were not able to discriminate glucose from γ-cyclodextrin at a significant level (proportion correct = 0.37, d’ = 0.728, P = 0.18 for low concentration; proportion correct = 0.41, d’ = 0.948, P = 0.08 for high concentration). Of those that correctly discriminated the low concentrations (9 participants), 3 correctly discriminated them a second time. Similarly, 10 participants correctly discriminated the high concentration γ-cyclodextrin from glucose and of those, 3 correctly discriminated the stimuli a second time (proportion correct = 0.30). This suggests that most subjects cannot differentiate the taste of γ-cyclodextrin from that of glucose.

Fig. 4.

Discriminability of glucose vs. γ-cyclodextrin at low (150 mM glucose, 75 mM γ-cyclodextrin) and high (280 mM glucose, 120 mM γ-cyclodextrin) concentrations. Subjects performed triangle tests to identify the target sample (γ-cyclodextrin) in a set of 3. The gray horizontal line indicates significant discriminability (P < 0.05).

General discussion

Cyclodextrins elicit taste

The present data show that α-, β-, and γ-cyclodextrins can be tasted at a significant level (P < 0.05) when delivered by themselves, and that their detection rates are equivalent to that of glucose at the same concentration (i.e. 75 mM; see Fig. 3, white bars). It is important to note that the functional concentration of β-cyclodextrin tested was approximately 18.5 mM due to its low solubility (Chatjigakis et al. 1992). In other words, the discriminability of β-cyclodextrin at ~18.5 mM is comparable to glucose, α- and γ-cyclodextrins at 75 mM. This finding itself is new given that taste perception of cyclodextrins has been mostly ignored in the past despite their wide usages in food, pharmaceutical, and consumer products (Biwer et al. 2002; Marcolino et al. 2011; Morrison et al. 2013).

Cyclodextrins are sweet stimuli

This study also provides the first psychophysical evidence suggesting that α-, β-, and γ-cyclodextrins are ligands of the human sweet taste receptor, hT1R2/hT1R3. The d’ values of the 3 cyclodextrins as well as glucose were significantly reduced (P < 0.05) in the presence of lactisole, a sweet taste blocker (see Fig 3, black bars). This finding implies that their taste detection is dependent on the sweet receptor. Cyclodextrins, in particular γ-cyclodextrin, have previously been described as slightly sweet (Szejtli and Szente 2005; Zakharova et al. 2016); however, these reports are not supported with data. Importantly, it is possible that the reported sweetness could have originated from smaller, sweet-tasting saccharides (e.g. maltose, maltotriose) that are produced from oral hydrolysis via salivary α-amylase activity (Lumholdt et al. 2012; Lim and Pullicin 2019). While salivary α-amylase may be less effective to hydrolyze α-1, 4 glycosidic linkages of cyclodextrins compared to their linear counterparts (Kondo et al. 1990; Kurkov and Loftsson 2013), amylases still are able to hydrolyze cyclodextrins (Jodái et al. 1984). In order to prevent such hydrolysis, the current study added acarbose, a salivary amylase inhibitor, to all test stimuli and corresponding water blanks. Accordingly, this study finding supports the conclusion that cyclodextrins are agonists of the canonical sweet receptor, and that their sweet taste is not due to oral hydrolysis byproducts.

The finding that cyclodextrins are sweet stimuli is somewhat surprising given their molecular size and structure. Simple sugars are relatively small molecules (e.g. fructose, glucose, and galactose: 180.16 g/mol; sucrose, lactose, and maltose: 342.30 g/mol). Cyclodextrins are much larger (972.85–1297.10 g/mol) than simple sugars. Nevertheless, the hT1R2/hT1R3 heterodimer is also known to be activated by certain sweet proteins (e.g. brazzein, thaumatin, monellin, miraculin) (Kant 2005), which are considerably larger (6.5–98.4 kDa) than cyclodextrins. Simple sugars typically activate the canonical sweet receptor via engulfment in a flexible binding pocket called the Venus Flytrap Domain (VFD) (Assadi-Porter et al. 2008; Masuda et al. 2012; Maillet et al. 2015). Sweet proteins bind to a different part of the receptor called the cysteine rich domain (CRD) (Kant 2005; Assadi-Porter et al. 2010). As this is the first report demonstrating cyclodextrins are sweet stimuli, further studies need to confirm that they are agonists of the sweet receptor and to determine how cyclodextrins interact with the receptor to activate it.

The low, but not significant, detectability, of γ-cyclodextrin in the presence of lactisole found in Experiment 1 remained a point of interest. While the sweetness of glucose, α- and β-cyclodextrins was completely blocked by lactisole (i.e. d’ = 0, proportion correct = 0.28–0.32; see Fig. 3), it was not the case for γ-cyclodextrin (d’ = 0.88, proportion correct = 0.40, p > 0.05). This suggests that approximately 7% of the subjects could possibly detect γ-cyclodextrin in the presence of lactisole. Note that the chance level of getting a correct answer for a triangle test is 33%. There are 2 possibilities to explain this result. First, it is possible that γ-cyclodextrin may elicit a side-taste other than sweetness. However, this is unlikely, as subjects did not report any side-tastes associated with γ-cyclodextrin. A more likely explanation is that γ-cyclodextrin is a sweet material, but its sweet taste may be incompletely blocked by lactisole. When subjects were asked to discriminate γ-cyclodextrin from glucose at 2 iso-intense concentrations in Experiment 2, the detectability did not reach a significant level at either concentration (p > 0.05; see Fig. 4). Importantly, those subjects who correctly discriminated γ-cyclodextrin from glucose reported that one stimulus was stronger than the other. Since individual differences in sweet taste sensitivity are well-established (Schiffman et al. 1985; Miller and Reedy 1990; Green et al. 1993; Cruz and Green 2000), it is reasonable to consider that the iso-intense concentrations we tested (low: 75 mM γ-cyclodextrin and 150 mM glucose; high: 120 mM γ-cyclodextrin and 280 mM glucose) were not an exact match for all subjects who participated in the study. Recent work has suggested the existence of an alternative oral glucose sensing mechanism (Yasumatsu et al. 2020; Breslin et al. 2021), which may contribute to intensity differences in glucose perception.

We currently hypothesize that γ-cyclodextrin may form inclusion complexes with lactisole. Previous research has shown that γ-cyclodextrin, but not α- and β-cyclodextrins, effectively sequesters gymnemic acid, another sweet taste blocker (Izutani et al. 2005). Gymnemic acid is a large molecule, with a molecular weight of 807 g/mol, and was previously determined to interact specifically with γ-cyclodextrin, likely because gymnemic acid is too large to fit in the binding pockets of α- or β-cyclodextrin. Γ-cyclodextrin may also form complexes with lactisole, which in turn limits lactisole’s ability to inhibit the sweetness of γ-cyclodextrin. In general, β- and γ-cyclodextrin are similarly effective at forming inclusion complexes, while α-cyclodextrin is typically ineffective (Ninomiya et al. 1998; Izutani et al. 2005; Przybyla et al. 2020). This is due to the size of the hydrophobic cavity formed by each cyclodextrin: α- is often too small (at 4.7–5.3 Å) to form inclusion complexes with most guest molecules, while β- and γ- (cavity sizes of 6.0–6.5 and 7.5–8.3 Å, respectively) can include molecules of varying size and hydrophobicity. Lactisole (molecular weight: 218.19 g/mol) is a smaller molecule than gymnemic acid, suggesting that it may fit in the binding pockets of both β- and γ-cyclodextrin. However, β-cyclodextrin has a low solubility (Chatjigakis et al. 1992), which in turn could have affected its ability to form inclusion complexes with lactisole at the concentration tested.

Chemical structures of oligosaccharides impact their taste perception

Oligosaccharides are a group of stimuli with diverse chemical structures and are prevalent in the human diet. These compounds can differ in several ways: base monomer (e.g. glucose, fructose, galactose), glycosidic linkages (e.g. α- or β-configuration; 1,4 or 1,6 bonds), chain length, terminal residues (reducing vs. nonreducing ends), or spatial orientation (e.g. linear, branched, cyclic). While cyclodextrins seem to be agonists of the canonical sweet receptor, their linear counterparts, MOS, are not; they are described as “starchy.” Our previous studies have shown that subjects could still detect MOS greater than DP 3, when the sweet taste receptor was blocked by lactisole (Lapis et al. 2016; Pullicin et al. 2017). Both MOS and cyclodextrin are made of glucose base monomers linked by α-1,4 glycosidic bonds (Szejtli 1998) and the chain-lengths of cyclodextrins are equivalent to the MOSs that were tested in the reported studies (Lapis et al. 2016; Pullicin et al. 2017). In addition, their detectability is equivalent at 75 mM: the d’ values of cyclodextrins reported in this study range between 1.77 and 2.39, while the d’ value of MOS DP 6-7 preparation was reported as 2.78 (Pullicin et al. 2017).

One unique difference between cyclodextrins and MOS is their spatial orientations: cyclodextrins are circular and occupy a roughly cylindrical shape in solution. MOS are linear chains of glucose that form a relaxed helix in solution; a glucose on one end of the chain can open to act as a reducing sugar. Because cyclodextrins are circular, their glucose monomers cannot open and therefore they do not possess a reducing end (see structures in Fig. 1). This suggests that the structure of these molecules dictates their interaction with their target receptor(s). However, it is currently unknown what chemical features of oligosaccharides are important to their taste perception. Future work should systematically test the impacts of various structural characteristics (e.g. glycosidic linkages, spatial orientation) on human taste perception of saccharides. Overall, this work underscores the critical importance of considering chemical structure when trying to understand chemical senses.

Funding

This research was supported by grant R01DC017555 from the NIH/National Institute on Deafness and Other Communication Disorders (NIDCD).

Conflict of Interest

The authors declare no competing financial interests. The funding agency was not involved in the conception, design, or execution of the study or the decision to publish study findings.