Short-term pectin-enriched smoothie consumption has beneficial effects on the gut microbiota of low-fiber consumers

Abstract Adequate consumption of fiber has a positive effect on health. The crossover study examined the effect of a pectin-enriched smoothie on gut microbiota and health parameters. During 3 weeks, 31 adults consumed two smoothies (11.6 or 4.8 g of fiber/day), alternating with washout periods in different order. At the end of each period, weekly food diaries, blood samples, and stool microbiota were collected. Changes in the microbiota during smoothie consumption were associated with baseline fiber intake. A greater proportion of up- (Lachnospira, Colidextribacter, and Bacteroides) or down-shifts (Streptococcus, Holdemanella) was observed in low-fiber (n = 22) compared to high-fiber consumers (n = 9). In both groups, the pectin-enriched smoothie reduced the number of the Ruminococcus torques group bacteria. Our results showed that the short-term approach is effective to estimate relationships between food components and gut bacteria.


Introduction
Human gut microbiota consists of about 10 13 microbial cells that inhabit mainly in the large intestine (Sender et al. 2016a ,b ).The diversity and richness of gut microbiota have a remarkable impact on human health through the production of different metabolites , vitamins , amino acids , and signaling molecules .T he composition of microbiota is affected to a large extent by host genotype , diseases , en vir onmental c hanges , and diet (T hursby and J uge 2017 ).It is supposed that e v en a very short-term, altered diet can modulate the human gut micr obiome r a pidl y.For example, Prevotella abundance decreases in vegetarian subjects during a fiveday long animal-based diet (David et al. 2014 ).Ho w e v er, it seems that the animal-based diet has a greater impact on the gut microbiota than the plant-based diet.It has been described that an animal-based diet increases bile acid-toler ant micr oor ganisms like Alistipes , Bilophila , and Bacteroides and decreases the abundance of bacteria that metabolize dietary plant pol ysacc harides-Roseburia , Eubacterium rectale , and Ruminococcus bromii (David et al. 2014 ).
Dietary fiber (DF) is a type of carbohydrate that is mostly not digestible by enzymes produced by humans.DF is transported thr ough the digestiv e system to the lar ge intestine, wher e it is metabolized by gut microbiota.The most abundant products of DF metabolism are short-chain fatty acids (SCFA) like butyrate, acetate, and propionate (Cummings 1981 ).These SCFAs differ in their fate and tissue distribution and can affect lipid, glucose, and c holester ol metabolism (Den Besten et al. 2013, Chambers et al. 2018 ).
DF can be divided into different groups-non-digestible oligosacc harides (inulin), non-starc h pol ysacc harides (pectin, beta-glucan), and resistant starch (Cummings and Macfarlane 1991 ).Pectin is one of the DFs that has se v er al beneficial effects on human health, including reduced glucose and c holester ol absorption (Stephen et al. 2017, Salleh et al. 2019 ); an increase of fecal mass (Cummings et al. 1978 ); and as a substrate for gut bacteria (Blanco-Pérez et al. 2021 ).T hus , citrus pectin oligosaccharides have been shown to have a modulating effect on c holester ol metabolism involving specific groups of bacteria along with their metabolites (Hu et al. 2019 ).Beta-glucan, known for its ability to enhance bile acid production from cholesterol in the liver and promote its ele v ated excr etion in the feces, is also linked to beneficial health effects.It has been reported that beta-glucan supplementation causes reductions in cholesterol and triglycerides (Cronin et al. 2021 ).Also, some studies show that psyllium may impr ov e the lipid profile (Deng et al. 2022 ).
Consuming a diverse range of fiber-rich foods is frequently linked to enhanced bo w el movement, lo w er body mass index (BMI), and impr ov ed glucose and c holester ol le v els, all of whic h ar e corr elated with a decr eased risk of cardiov ascular disease, obesity, and type 2 diabetes (Reynolds et al. 2019 ).A systematic r e vie w and meta-analysis of 12 randomized controlled trials suggests that supplementation with isolated soluble fiber can impr ov e body composition and metabolic outcomes, including fasting glucose and insulin le v els (Thompson et al. 2017 ).The recommended daily fiber intak e de pends on dail y ener gy consumption and is about 13 g per 1000 kcal (Pitsi et al. 2017 , Stephen et Figur e 1. T he study was divided into 5 periods: a base period to describe the individual health parameters and food habits at a starting point, two different smoothie consumption periods (with high and low pectin content), followed by smoothie WO periods.WO periods pr efigur e a regular menu and no smoothie consumption; after that, participants were grouped based on their daily fiber intake.Low-fiber intake group (LF) consumed < 23 g of DF per day ( n = 22), high-fiber intake group (HF) consumed > 23 g of fiber per day ( n = 9).al. 2017 ).Ho w e v er, the r eal amount of DF consumption is usuall y m uc h lo w er for people follo wing a Western-type diet enriched with products of animal origin (Stephen et al. 2017 ).
To date, a complex a ppr oac h that includes k ee ping a food diary, measuring general body parameters, conducting extensive blood tests, and continuousl y c har acterizing gut micr obiota in r elation to two different levels of increased fiber consumption and following washing-time periods has not been performed.The present study aimed to e v aluate the effect of a smoothie enriched with pectin and other soluble fibers on human gut microbiota and health parameters depending on the daily fiber intake during a short-term intervention study.The results of this study provide new insight into the support of beneficial bacteria by increased intake of soluble DFs.

Materials and methods
The current study was carried out and the smoothies were developed by the Center of Food and Fermentation Technologies (TF-TAK, Estonia) in collaboration with Siidrikoda (Estonia).

Composition of the smoothies
In the present study, we used two different fruit smoothies with distinctive fiber (esp.pectin) content.The smoothies consisted mainly of fruit purees and concentrates .T he high-pectin (HPect) smoothie recipe was developed with the aim of utilizing apple pomace from the cider production facility Siidrikoda, thereby adding value to the byproduct.It contained 36% of apple puree and a special dietary supplement included pectin, psyllium, and betaglucan fibers (Calm your rumbly tummy, Elsa vie , Estonia).T he lo w-pectin smoothie (LP ect) contained an a ppr opriate amount of starch.The exact recipes of the smoothies are presented in the Supplementary Table S1 , and the nutritional values of the smoothies per 100 g are shown in the Supplementary Table S2 .

Recruitment of study participants and design of the study
The current study was carried out from October 2022 to December 2022.Participants were selected according to a questionnaire and definite criteria.The questions concerned the respondents' eating habits, description of their health status (incl.digestive health), and data on their general lifestyle .T he tar get gr oup of the study included individuals who met at least 3 of the following conditions: sedentary lifestyle, low Bristol score ( < 3), modest consumption of fruits and vegetables, and gastrointestinal com-plaints without a diagnosed disease .T he exclusion criteria were as follows: antibiotic use 3 months before the study, any severe or c hr onic diseases (e.g.cancer, Cr ohn's disease, ulcer ativ e colitis, etc.), specific diet (e .g. ketogenic , v egan, low carbohydr ate, high-fat diet), tr av el to subtropic or tropic area within 3 months, or intolerance of a pples.P articipants wer e asked to follow their usual eating habits throughout the study.Subjects who did not follow the study pr otocol wer e excluded fr om the study.Out of 368 volunteers who completed the questionnaire, 55 were men, representing 15% of the total participants.Out of the 39 eligible participants, 31 completed the study, of which 3 (9.7%)were men .
The study consisted of five periods (Fig. 1 ).The one-week base period was followed by two three-week test periods with alternating tw o-w eek w ashout (WO) periods.Test periods encompass tw o different smoothie consumption (one per period) in addition to the regular menu.
Following the base period, subjects were randomly divided into two groups based on the first smoothie they ate, while the study participants were unaware of the type of smoothie they were consuming.Group 1 ( n = 17) consumed a HPect smoothie first, and group 2 ( n = 14) started with an LPect smoothie.After the WO period, the smoothies were switched between the groups (Fig. 1 ).
For data analysis, subjects were grouped into low-fiber (LF) and high-fiber (HF) intake cohorts based on the analysis of food diaries according to the Estonian nutrition and movement recommendations (Pitsi et al. 2017 ) and their e v eryday DF consumption (during base and WO periods).The LF intake group ( n = 22) regularly consumed < 23 g of DF per day, while the HF intake group ( n = 9) consumed > 23 g of DF per day (Fig. 1 ).All participants signed written informed consent forms before the beginning of the study.The r esearc h was a ppr ov ed by the local ethics committee (Research Ethics Committee of the National Institute for Health De v elopment, Reference number 1065, issued on 04/25/2022).

Data collection
Nutrition data diaries were filled out one week before blood and fecal sample collection using the NutriData dietary analysis program (National Institute for Health Development, Estonia).The participants were asked to fill out the exact amount of smoothies they consumed.The analysis of macronutrients considered all regular diet periods (base and WO periods) to rule out occasional variations in the daily diet.At the end of every period, blood and fecal samples were taken (five blood and fecal samples per participant), and body composition was determined with a Tanita body composition analyzer (DC-360S, Tanita Corporation, Tok y o, J apan).

Analyses of blood samples
Blood test analysis was performed by the Tartu University Hospital laboratory or SYNLAB Eesti OÜ (Tallinn, Estonia).Blood samples were taken between 7 a.m. and 10.30 a.m. after overnight fasting (on the same day when the fecal sample was produced).Blood was centrifuged and serum was used for analysis .T he hexokinase method was used to determine glucose le v el.Enzymatic colorimetric method was used to measure blood lipids (total, LDL, HDL-c holester ol, and trigl ycerides), gamma glutamyl transferase (GGT), and uric acid levels (Cobas c 501, Roche Diagnostics).Serum alanine aminotr ansfer ase (ALAT) le v el was tested using a kinetic photometric method (Cobas 6000, Roche Diagnostics CH-6343 Rotkr euz, Switzerland).Imm unoturbidimetric test was used to measure C-reactive protein level.An automatic analyzer was used to determine the hemogram together with the leukogram.

DN A extr action and sequencing
Fecal samples were collected with DNA/RNA Shield Collection Tubes with Swabs (Zymo Research, Irvine , C A, USA) using F e-cesCatcher by TagHemi (Zeijen, The Netherlands) and stored at + 4 • C. Before DNA extraction, samples were frozen at −20 • C at least overnight.DN A w as extracted using the ZymoBIOMICS DNA Miniprep Kit (Zymo Research, Irvine, CA, USA) according to the manufacturer´s instructions.Qubit™ 3 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and the dsDNA BR Assay Kit (Thermo Fisher Scientific) were used for gDNA quantification.
The V4 hypervariable region of the 16S rRNA gene was PCR amplified using universal forward F515 5`-GTGCCAGCMGCCGCGGTAA-3`and r e v erse R806 5´-GGA CTA CHV GGGTWTCTAAT-3´primers (Ca por aso et al. 2011 ).Samples were sequenced using the Illumina MiSeq platform and a 2 × 150-cycles paired-end sequencing protocol.On aver a ge, 38 800 r eads (minim um 17 677 r eads) per sample wer e obtained.The whole sequencing w orkflo w w as published before (Kazantse v a et al. 2021 ).
DNA sequence data was analyzed by BION-meta software ( https:// github.com/nielsl/ mcdonald-et-al ) according to the authors' instructions (McDonald et al. 2016 ).The sequences were first cleaned at both ends using a 99.5% minimum quality threshold for at least 18 out of 20 bases for 5`-end and 28 out of 30 bases for 3`-end, follo w ed b y joining and r emov al of shorter contigs than 150 bp.Afterw ar ds, the sequences w er e cleaned fr om c himer as and clustered by 95% oligonucleotide similarity (k-mer length of 8 bp, step size 2 bp).Finally, consensus reads were aligned to the SILVA r efer ence 16S rRNA database (v138) using a w or d length of 8 and a similarity cut-off of 90%.

Sta tistical anal yses
Statistical analyses were performed at the bacteria genus le v el with abundance at least > 0.0006 per sample.Data analyses were done by R version 4.2.1 (The R Foundation for Statistical Computing, Vienna, Austria) using open public pac ka ges-indicspecies, dplyr , ggpubr , reshape2, tidyverse, and ggplot2, grDevices package were used for visualization.
Pairwise comparisons were evaluated using the Wilcoxon signed-rank test.All pairwise comparisons were calculated between the baseline and the end of the smoothie consumption or W O period.The W ilcoxon rank-sum test was used to e v aluate differences in consumption of macronutrients and abundances of bacterial genera between LF and HF intake groups.
Statistical analysis of general health parameters was performed using the SPSS for Windows version 20.0 (SPSS Inc. Chica go, IL, USA).Nonpar ametric Wilcoxon or Mann-Whitney Utests were used to e v aluate the differ ence between the medians and distributions of the outcome parameters for the subgroups.Statistical significance was defined at P < 0.05.

The effects of short-term smoothie consumption on gut microbiota are temporary
To exclude the influence of seasonal eating habits on the effect of target smoothie consumption, we used the cr ossov er study design, dividing participants into two groups based on the order of smoothies intake (Fig. 1 ).We first examined the changes in the abundance of 177 bacterial genera represented in the study gr oup during eac h of the fiv e periods.We compar ed samples after the smoothie consumption periods with samples from the presmoothie consumption periods .T he analysis confirmed that micr obiota c hanges during the first smoothie period did not affect the results of the second smoothie period.Most of the effects caused by smoothie consumption were provisional and did not show differences according to the order of smoothie consumption among the pr e v alent taxa (Fig. 2 ).The abundance of Dorea decreased after HPect period in group 1, a similar tendency was observed in group 2, but not statistically significantly.The abundances of R .torques group and Coprococcus decreased after HPect period in both gr oups.Furthermor e, LPect smoothie decr eased the le v el of Coprococcus in group 1.The HPect had a remarkable positive effect on the proportion of Lachnospira in group 1, and a slight but not significant increase was detected in group 2 (Fig. 2 ).By the end of the WO periods, the abundances of all these genera returned to the base le v els, confirming that the applied crossover method is suitable to e v aluate the effect of fiber-enriched foods on the gut microbiota.

T he char acteristics of tw o different fiber intake groups during regular diet period
Consumption of macronutrients based on food records during a regular diet was analyzed in two fiber intake groups.During the study, we formed two focus groups based on the received data.Despite both groups having a low consumption of fruits and vegetables based on the food frequency questionnaire prior to the study, 9 participants had sufficient fiber consumption during the smoothie trial.The LF group consumed fewer fruits and vegetables, as well as other HF foods like whole grains , legumes , and nuts.In contrast, the HF group mainly obtained their sufficient daily fiber intake from whole grain products , legumes , and nuts , with a lesser contribution from fruits and vegetables.While the HF gr oup serv ed as a contr ol, the LF consumption gr oup had more participants as it was the main point of interest.In comparison to fiber intake, the other macro-nutritional values in the dail y menu, suc h as ener gy, carbohydr ates, fat, and protein consumption did not show any significant difference between the two groups (Table 1 ).Ho w ever, it is w orth noting that fat consumption for both groups was higher than recommended in Estonian nutrition guidelines (Pitsi et al. 2017 ).T hus , 40% and 36% of the daily consumed energy come from fat in the LF and HF intake groups, r espectiv el y.
To investigate the differences in gut microbiota between these tw o groups, w e compared the ten most abundant bacteria genera in the LF and HF intake groups at the beginning of the study.We found that the groups had quite similar gut microbiota composition at the genus le v el.Ho w e v er, Pr e votella-9 dominated ov er  Data ar e pr esented as av er a ge ( ±stde v).Macr onutrients ar e pr esented as a percenta ge fr om dail y ener gy expenditur e of r egular diet periods (no smoothie consumption).Distribution was compared between the groups using the Wilcoxon rank-sum test.Statistically significant differences ( P < 0.05) are shown by asterisk ( * ).LF-low-fiber intake group ( n = 22); HF-high-fiber intake group ( n = 9).
Bacteroides in the LF intake group, whereas the opposite trend was observed in the HF intake group (41% and 22% of the participants had higher le v el of Pr e votella-9 than Bacter oides in the LF and HF intake gr oups, r espectiv el y) (Fig. 3 A).At a lo w er abundance le v el, certain bacterial gener a exhibit statistically significant differences between LF and HF intake groups (Fig. 3 B).We observed that participants who consume < 23 g of DF per day (LF intake group) exhibit a decreased abundance of Coprococcus , Akkermansia, Lactococcus , and Cloacibacillus as compared to the HF intake gr oup ( P -v alues 0.046, 0.036, 0.031, and 0.008, respectiv el y).

The gut microbiota of LF intake group is more affected by smoothie consumption than HF intake group
To understand how short-term consumption of regular and pectin-enriched fruit smoothies can affect the human microbiota, we analyzed samples from two groups with varying basal fiber intake, as described in Table 1 .Regarding the HPect, the abundances of Lachnospira and Bacteroides incr eased, wher eas the abundances of Subdoligranulum, Coprococcus, E .fissicatena group , Dorea, Lactococ-cus, Evtepia, and Enterococcus decreased after consumption (Fig. 4 A, Supplementary Table S3 and S4 ).It is worth noting that the baseline abundances of Coprococcus and Lactococcus were higher for the HF intake group (Fig. 3 B).In the LF intake group, the most significant increase in prevalence was detected for Coldextribacter after the LPect period .The higher abundance of CAG-352 and Oxalobacter was also observed in the LF intake group post-LPect consumption.On the other hand, the abundances of Turicibacter, Peptococcus, Defluviitaleaceae_UCG-011, UCG-009 , and Merdibacter wer e r educed after LPect consumption in this gr oup (Fig. 4 B, Supplementary Table S3 and S4 ).Furthermore, the prevalence of Streptococcus and Holdemanella decreased after both smoothie consumption periods (Fig. 4 C, Supplementary Table S3 and S4 ).
In the HF intake gr oup, ther e wer e significantl y fe wer changes after the smoothie periods compared to the LF intake group.In this group, the abundances of Lachnoclostridium, Lachnospir aceae_NK4A136_group , and Par asutterella wer e incr eased, and Anaersotipes le v el r educed after HPect consumption (Fig. 5 A).LPect smoothie increased the abundance of Bilophila and decreased the proportion of Agathobacter (Fig. 5 B).Lo w er le v els of Ruminococcus were detected after both smoothie periods in the HF intake group, but no significant shifts in the abundances of these genera were observed in the LF intake group (Fig. 5 C, Supplementary Table S3  and S4 ).
Ther e wer e se v er al gener a that exhibited significant c hanges in both groups after the consumption of smoothie .T he abundance of Oscillibacter increased after the LPect period in both groups .T he pr e v alence of Erysipelotrichaceae_UCG-003 decreased after LPect consumption in the LF intake group and after the HPect period in the HF intake group.The abundance of R .torques group decreased after the HPect period in both groups .T he level of E .ventriosum group decreased following both smoothie periods in the LF intake group and after LPect consumption in the HF intake gr oup.Finall y, the abundance of Paraprevotella increased after both smoothie consumption periods in the LF intake group and after LPect in the HF intake group.(Fig. 6 , Supplementary Table S3  and S4 ).

HF intake is associated with lower body weight and better health parameters
The compar ativ e health indicators of the LF and HF intak e stud y groups during the base period are presented in Supplementary Table S5 .The median BMI of subjects in the HF intake group was statistically significantly lo w er compared to the median BMI of the LF intake group (20.6 vs. 23.0,Mann-Whitney U test, P = 0.048).In addition, compared to the LF intake group, the HF intake group sho w ed slightly lo w er mean blood glucose (5.1 vs. 4.8 mmol/l, Mann-Whitney U test, P = 0.048) and hemoglobin levels (134.5 vs. 130 g/l, Mann-Whitney U test, P = 0.048).Because men have a higher range of r efer ence v alues for uric acid and hemoglobin, these data wer e anal yzed separ atel y for women.By excluding three men from the analysis, the statistical difference in hemoglobin le v els between gr oups disa ppear ed.Ho w e v er, a statisticall y significantl y higher le v el of uric acid was r e v ealed in the LF intake group ( n = 20) compared to the HF intake group ( n = 8) (261.0 vs. 229.5 μmol/l, Mann-Whitney U test, P = 0.03).No differences were found between other health parameters in the baseline period.Baseline differences were maintained throughout the study periods.No significant changes in health parameters were obtained with the use of smoothies in the LF intake group except for glucose, where the between-group difference disappeared by the end of the study.

Discussion
P articipants whose e v eryday fiber intake w as lo w er than 23 g per day had m uc h mor e alter ed gut micr obiota gener a in r esponse to increased fiber intake .Moreo ver, these changes are directed and encompass distinct genera.It has been established that a person's   dietary habits can influence the changes of gut microbiota, and whether they respond to a particular treatment or not (Rose et al. 2015, Tap et al. 2015, Pr oc házk ová et al. 2023 ).In this study, we sho w ed that the effects of regular and DF-enriched fruit smoothies depend on the baseline consumption of fiber.T hus , we ma y expect some substantial positive effect of DF on gut microbiota in case of low baseline fiber consumption.Howe v er, these c hanges are short-term and disappear after restriction of increased fiber intake.
The most abundant bacteria did not show significant changes betw een these groups; ho w ever, the ratio of the abundances of Prevotella 9 and Bacteroides seems to be different (41% and 22% of the participants have a higher abundance of Prevotella 9 than Bacteroides in LF intake and HF intake gr oup, r espectiv el y) .Pr e viously, it has been known that Prevotella should dominate for people who consume more vegetarian food rich in DF, and Bacteroides is more common for people who prefer animal-based food rich in fat and protein (Wu et al. 2011, Roager et al. 2014 ).Differing results in our study may be caused by the fact that the landscape of Prevotella impact on gut microbiota is diverse and cannot be associated with increased fiber consumption only.For instance, healthy subjects who exhibited impr ov ed glucose metabolism after kernel-based bread intake had a higher Prevotella/Bacteroides ratio (K o vatcheva-Datchary et al. 2015 ) or lost more body weight and fat compared to individuals with low Prevotella/Bacteroides in response to a diet rich in fiber (Hjorth et al. 2019 ).Also, a habitual diet rich in fibers can promote the growth of different Prevotella strains (De Filippis et al. 2019 ), some of which do not have a positive impact on human health, and their increased levels could be linked with inflammation (Larsen 2017 ).It is important to mention that both Prevotella and Bacteroides are the primary pectin consumers and could equally benefit from a HF smoothie (Larsen et al. 2019 ).
On the other hand, we observ ed distinctiv e c hanges in the proportion of four genera between groups consuming various amounts of fiber.Participants who consumed < 23 g of DF per day exhibited a decreased abundance of Coprococcus , Akkermansia, Lactococcus , and Cloacibacillus as compared to the HF intake group.It was shown that increased levels of these bacterial genera in the gut relate to better health parameters and digestive activity.Some species from the genus Coprococcus are butyrate producers and thereby could have a supportive effect on human health and inhibit inflammatory processes in the colon (Rivière et al. 2016 ).
Coprococcus can also participate in gut-br ain comm unications and has a positiv e corr elation with a better mental and emotional state (Valles-Colomer et al. 2019 ).There is some evidence that shows an association of Coprococcus comes and higher BMI (Liu et al. 2017, Adamberg et al. 2020 ).Conversely , in our study , the HF intake group had a higher Coprococcus and a lo w er BMI than that in LF intake group.Ho w ever, it can be explained by the different species compositions of coprococci that we could not discriminate this time.Also, the abundance of Akkermansia was higher in the HF intake group.Most of the cases, Akkermansia , especially Akkermansia muciniphila is known as a good health marker due to the ability to r ene w the m ucus layer and ther eby hav e a supportiv e effect on maintaining the intestinal integrity.Many diseases are associated with a decreased level or absence of Akkermansia (Geerlings et al. 2018 ).At the same time , we ha ve to pay attention that a high le v el of the abundance of Akkermansia is associated with constipation and longer gut transit time (Vandeputte et al. 2016, Asnicar et al. 2021 ).Since a sedentary lifestyle and a low Bristol scor e wer e inclusion criteria for the study, this may explain why some participants have very high levels of Akkermansia .
We found that the abundance of Lac hnospira incr eased after HPect consumption in LF intake group, while the abundance remained unchanged in HF intake group.It could be assumed that the effect we observed after HPect period is associated with pectin digestion.HPect contains a notable amount of pectin and individuals who consume a lot of fruits have more Lachnospira in their gut (Kable et al. 2022 ).Lachnospira can use pectin as a substrate for acetate production (Bang et al. 2018, Beukema et al. 2020 ) and could contribute to the formation of acidic environment in the colon, which in turn impedes the growth of pathogenic bacteria.Moreov er, the genus Lac hnospira has a higher abundance among lean people, and a negative correlation with total and LDL cholesterol le v els is described (Companys et al. 2021 ).
One more positive effect of increased fiber consumption was the reduction of mucus-degrading R .torques group after HPect intake for both groups .T he proportion of this bacterium is usually higher for individuals with intestinal inflammation.It has been shown that the m ucol ytic R .tor ques is markedl y incr eased in noninflamed UC (ulcer ativ e colitis) and Cr ohn's disease epithelium (Png et al. 2010 ).Ther efor e, we should note that, besides considering the daily quantity of DF consumption, it is essential to pay attention to the type of consumed DF.
Regarding the health risks, we confirmed an already known association of higher fiber intake with lo w er BMI (Hadrévi et al. 2017, Ito et al. 2023 ).In ad dition, the stud y sho ws that w omen in the LF intake group have higher uric acid levels compared to the HF intake group.LF Western diet is associated with higher levels of uric acid, whic h incr eases the risk of gout.The association of a HF diet with lo w er uric acid le v els in women has also been shown previously (Chen et al. 2020 ).Although we observed between-group differences in BMI and health parameters at baseline, short-term increased fiber intake was not sufficient to impr ov e health outcomes in either study group.
It is worth mentioning that people tend to a ppr eciate their eating habits not v ery adequatel y, whic h limited the study by decreasing the number of recruited participants based on the questionnaire .Despite not ha ving healthy food habits , according to the questionnair e, nearl y a thir d of the study participants w ere in good health.
A weakness of the study is that few men participated, and men also dropped out more easily, resulting in a biased sample to w ar ds women.Another important aspect to consider is that the small group is susceptible to individual variability, and each sample can significantly impact the overall results.

Conclusion
In our study, we demonstrated that industrial by-products such as apple pomace could be used as a main component, along with other fibers, to create snacks with health benefits that enrich the regular diet and increase fiber consumption.Using a cross-section study, in which participants filled dietary questionnaires, tested various human health parameters, took blood tests, and determined changes in gut micr obiota.We r e v ealed that baseline fiber intake determines the response of gut microbiota to smoothie consumption.Although smoothie consumption did not provide benefits on the blood markers, we observed that LF consumers had more positive changes in their microbiota at the genus level after smoothie intake .Hence , a pectin-enriched smoothie could promote the growth of beneficial gut bacteria.Moreover, we estimated the correlation between food components and definite gut bacteria.Ho w e v er, this effect is transient, as most changes in micr obiota r ecov er ed during WO time.Certainl y, for e v aluating the

Figure 2 .
Figure 2. Bacterial up-and down-shifts after smoothie periods and r ecov eries of abundances after WO periods.Box plots are arranged in period order.Base period c har acterizes micr obiota befor e the interv ention.During the interv ention, gr oup 1 consumed first pectin-enric hed smoothie (HPect), follo w ed b y w ashout (WO HP ect), and consumption of control smoothie (LP ect) with w ashout (WO LP ect).In the gr oup 2 interv ention was started with the LPect smoothie.Each dot represents one individual, and the same color was used for one person in different periods in the group.* P ≤ 0.05, * * P ≤ 0.01, ns-not significant.Y-axis-abundance, normalized to 1.

Figure 3 .
Figure 3. (A) Heatmap illustrating the abundances of the most abundant genera in the base period of LF and HF intake group members, based on the daily fiber intake.(B) At lo w er levels of abundance, four genera show the differences between LF and HF intake groups in the base period.Dots r epr esent the individuals in the group .W ilcoxon rank-sum tests between LF and HF intake groups were performed ( * P ≤ 0.05, * * P ≤ 0.01), Y-axis-abundance, normalized to 1. LF-low-fiber; HF-high-fiber.

Figure 4 .
Figure 4. Alter ed gener a in LF intake gr oup .The pairwise comparison was performed using the W ilcoxon signed-r ank test for statistical anal ysis to compare LPect and HPect effects on base period microbiota at the genus level ( * P ≤ 0.05, * * P ≤ 0.01, * * * P ≤ 0.001, ns-not significant).Each dot r epr esents one individual, and the same color was used for one person in different periods in the group.Y-axis-abundance, normalized to 1. (A) HPect alter ed gener a; (B) LPect alter ed gener a; (C) Gener a alter ed b y both smoothies.LP ect-lo w pectin smoothie; HP ect-high pectin smoothie.

Figure 5 .
Figure 5. Alter ed gener a in the HF intake gr oup .The pairwise comparison was performed using the W ilcoxon signed-r ank test for statistical anal ysis to compare LPect and HPect effects on base period microbiota at the genus level ( * P ≤ 0.05, * * P ≤ 0.01, ns-not significant).Each dot represents one individual, and the same color was used for one person in different periods in the group.Y-axis-abundance, normalized to 1. (A) HPect altered genera, (B) LPect altered genera, and (C) Genera altered by both smoothies.LPect-low pectin smoothie; HPect-high pectin smoothie.

Figure 6 .
Figure 6.Alter ed gener a in both gr oups .T he pairwise comparison was performed using the Wilcoxon signed-r ank test for statistical anal ysis to compare LPect and HPect effects on base period microbiota at the genus level ( * P ≤ 0.05, * * P ≤ 0.01, ns-not significant).Each dot represents one individual, and the same color was used for one person in different periods in the group.Y-axis-abundance, normalized to 1. (A) LF intake group; (B) HF intake group.LP ect-lo w pectin smoothie; HPect-high pectin smoothie.

Table 1 .
The consumption of macronutrients during a regular diet.