Health benefit of vegetable/fruit juice-based diet: Role of microbiome

Abstract

The gut microbiota is an important contributor to human health. Vegetable/fruit juices provide polyphenols, oligosaccharides, fiber and nitrate (beet juice), which may induce a prebiotic-like effect. Juice-based diets are becoming popular. However, there is a lack of scientific evidence of their health benefits. It was our hypothesis that changes in the intestinal microbiota induced by a juice-based diet play an important role in their health benefits. Twenty healthy adults consumed only vegetable/fruit juices for 3 days followed by 14 days of customary diet. On day 4 we observed a significant decrease in weight and body mass index (p = 2.0E⁻⁰⁵), which was maintained until day 17 (p = 3.0E⁻⁰⁴). On day 4 the proportion of the phylum Firmicutes and Proteobacteria in stool was significantly decreased and Bacteroidetes and Cyanobacteria was increased compared to baseline and was partially reversed on day 17. On day 4 plasma and urine nitric oxide was increased by 244 ± 89% and 450 ± 360%, respectively, and urinary lipid peroxidation marker malondialdehyde was decreased by 32 ± 21% compared to baseline. General well-being score was increased at the end of the study. In summary a 3-day juice-based diet altered the intestinal microbiota associated with weight loss, increase in the vasodilator NO, and decrease in lipid oxidation.

Full text

1
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
www.nature.com/scientificreports
Health benet of vegetable/
fruit juice-based diet: Role of
microbiome
Susanne M. Henning , Jieping Yang, Paul Shao, Ru-Po Lee, Jianjun Huang, Austin Ly, Mark
Hsu, Qing-Yi Lu, Gail Thames, David Heber & Zhaoping Li
The gut microbiota is an important contributor to human health. Vegetable/fruit juices provide
polyphenols, oligosaccharides, ber and nitrate (beet juice), which may induce a prebiotic-like eect.
Juice-based diets are becoming popular. However, there is a lack of scientic evidence of their health
benets. It was our hypothesis that changes in the intestinal microbiota induced by a juice-based diet
play an important role in their health benets. Twenty healthy adults consumed only vegetable/fruit
juices for 3 days followed by 14 days of customary diet. On day 4 we observed a signicant decrease in
weight and body mass index (p = 2.0E−05), which was maintained until day 17 (p = 3.0E−04). On day 4
the proportion of the phylum Firmicutes and Proteobacteria in stool was signicantly decreased and
Bacteroidetes and Cyanobacteria was increased compared to baseline and was partially reversed on day
17. On day 4 plasma and urine nitric oxide was increased by 244 ± 89% and 450 ± 360%, respectively,
and urinary lipid peroxidation marker malondialdehyde was decreased by 32 ± 21% compared to
baseline. General well-being score was increased at the end of the study. In summary a 3-day juice-
based diet altered the intestinal microbiota associated with weight loss, increase in the vasodilator NO,
and decrease in lipid oxidation.
Vegetable/fruit juice-based diets have been very popular recently. However, well designed controlled research
studies with clinical outcome measures providing scientic evidence of potential health benets of juice only
diets are limited1. e consumption of vegetable/fruit juice during the abstinence from food provides essential
nutrients and improves compliance.
Fruit and vegetables are rich sources of several biologically active components that contribute to general
health and decrease the risk of chronic diseases such as cardiovascular disease2. ey are the most ubiquitous
source of phenolic compounds3. Polyphenols exert a variety of physiological eects in vitro including antioxida-
tive, immunomodulatory and antimicrobial activities4.
e absorption of polyphenols in the small intestine is limited and considerable amounts of these polyphenols
can be found in the colon. ere the colonic bacteria metabolize polyphenols to smaller compounds, which in
turn alter the abundance of bacteria in the intestinal microbiome. In addition, fruit and vegetable are rich in fer-
mentable ber with prebiotic activity. High ber intake is associated with decreased risk of cardiovascular disease,
type 2 diabetes and some forms of cancer5. Fiber is composed of oligosaccharides, which resist digestion in the
small intestine and are transported to the colon where they provide energy for gut bacteria6. Growing evidence is
demonstrating the role of the microbiota in the health benets of dietary ber consumption6.
In the human intestine the gut microbiota is an important contributor to human health and has been impli-
cated in the development of obesity and obesity-related diseases such as diabetes and cardiovascular disease7–9.
e two most abundant bacterial phyla in humans and in mice are Firmicutes (40–60%) and Bacteroidetes (20–
40%) with lower abundance of Actinobacteria, Fusobacteria, Proteobacteria and Verrucomicrobia10. Recent stud-
ies show that dietary interventions with polyphenol rich extracts and foods, including dealcoholized red wine
polyphenols, cocoa-derived avanols, quercetin and grape anthocyanins, modulate the human gut microbiota by
decreasing the abundance of Firmicutes and increasing Bidobacteria, Lactobacillus and Verrucomicrobia11–13,
which is also a key dierence in the gut microbiota found in obese and lean individuals14, 15.
Center for Human Nutrition, David Geen School of Medicine, Department of Medicine, University of California
Los Angeles, California, CA, USA. Correspondence and requests for materials should be addressed to S.M.H. (email:
shenning@mednet.ucla.edu)
Received: 8 December 2016
Accepted: 6 April 2017
Published: xx xx xxxx
OPEN

www.nature.com/scientificreports/
2
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
We therefore investigated whether the consumption of fruit and vegetable juices (6 bottles daily of mixtures of
greens, roots, citrus, lemon, cayenne and vanilla almond) as part of a 3-day juice only program alters the intestinal
microbiota in twenty healthy participants. Secondary outcomes of the study were to determine the eect of the
3-day juice based diet on change in weight and body composition and biomarkers of oxidation (urine malondial-
dehyde) and vasodilation (plasma and urine nitric oxide).
Results
irty participants were screened. Twenty-ve participants met enrollment criteria and were randomized and
completed the 31-day study. Five participants were not able to provide stool samples at either day 4 or 17 and were
excluded. All data included in this manuscript includes the twenty participants only (Table1).
Body weight and composition. According to the calorie content provided by the manufacturer the total
calorie intake per person per day was 1310 kcal. During the 3-day juice intervention a signicant weight loss
of 1.7 ± 1.2 kg was observed (p = 2.0E−05) (Fig.1a). Aer the 2-week follow up period body weight remained
decreased (0.91 ± 0.9 kg) compared to baseline weight. Body mass index (BMI) was decreased by 0.6 ± 0.4 aer
the 3-day juice fast and remained decreased by 0.33 ± 0.3 aer the 2 week follow up period (Fig.1b).
Intestinal microbiome. Fecal microbiota composition of 20 participants collected at day 0, 4 and 17 was ana-
lyzed by sequencing the V4 region of the 16S rDNA gene. Rarefaction curves for bacterial DNA sequences for each
stool sample approached a plateau, indicating that sequence coverage was sucient to encompass the majority of
diversity contained within each sample (results not shown). e distribution of 16S rDNA genes at the phylum
level was composed of Firmicutes > Bacteroidetes > Proteobacteria = Verrucomicrobia > Actinobacteria (52 ± 4,
41 ± 4, 2.3 ± 1, 2.3 ± 1 and 5 ± 4%, respectively). e juice consumption, however, was associated with a signicant
decrease in the proportion of the bacterial phylum Firmicutes (p = 0.014) and increase in Bacteroidetes (p = 0.026)
and Cyanobacteria (0.003) at day 4 compared to baseline and was partially reversed to baseline proportions at day
17 (Fig.2). e proportions of Verrucomicrobia, Proteobacteria, Actinobateria and Fusobacteria were not changed
signicantly (Fig.2). Linear regression analysis of data from this study showed a signicant positive correlation with
day 4 body weight and Firmicutes proportion (p = 0.006), and a negative correlation with Bacteroidetes (p = 0.011)
(Supplementary Fig. S1). ere was no dierence in indices of fecal community richness and diversity (α-diversity
indices) throughout the intervention period (Supplementary Fig. 2; p = 0.49).
On the genus level the most prominent bacterium was Bacteroides, which was increased signicantly at day 4
compared to baseline (Table 2 and Supplementary Fig. S3). Twenty dierent Bacteroides species were identied
and the abundance of eight of them (B. acidifaciens, B. caccae, B. fragilis, B. massiliensis, B. nordii, B. salyersiae, B.
thetaiotaomicron, B. uniformis) was increased on day 4 compared to baseline (Supplementary Fig. S3).
On the genus level the following bacterial populations were also signicantly increased at day 4 compared to
baseline (percent of baseline): Halospirulina (1467%), Paraprevotella (348%), Barnesiella (200%), Odoribacter
(200%) and Bacteroides (144%) (Table2). On the other hand, proportions of the following bacterial genera were
decreased signicantly at day 4 compared to baseline (percent of baseline): Streptococcus (8%), Subdoligranulum
(30%), Eisenbergiella (40%), Ruminiclostridium (50%), and Dialister (67%) (Table2). Except for Streptococcus,
which was still signicantly decreased to 19% of baseline at day 17, the proportions of all other bacterial genera
were changed back to baseline values (Table2).
e functional analysis of the microbial metagenome using PICRUST showed no dierence in bacterial nitro-
gen metabolism (Supplementary Fig. S4).
Var iable
Study Participants
Screened (N) 30
Enrolled/randomized (N) 25
Participants with all stool samples (N) 20
Age, years 32 ± 8
Gender
Male(N) 5
Female(N) 15
Race
African American(N) 1
Caucasian(N) 9
Asian (N) 8
Bi-Racial(N) 2
Height, cm 167 ± 5
Weight, kg 71 ± 18
BMI, kg/m225.5 ± 5
Table 1. Baseline characteristics of the study participants.

www.nature.com/scientificreports/
3
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
Plasma antioxidant capacity and urine lipid peroxidation. e eect of juice based diet on plasma
antioxidant capacity was determined by analyzing the trolox equivalents (TE) using the TEAC method. ere was
no change in plasma TE comparing samples collected at day 4 and 17 with baseline samples (data not presented).
Lipid peroxidation was determined by the analysis of urine malondialdehyde (MDA). Urine MDA was signi-
cantly decreased by 40% in 6-hour urine collected at day 4 compared to baseline urine (p = 0.01) and returned to
baseline levels at day 17 (Fig.3).
Plasma and urine nitric oxide. To evaluate potential eects of vegetable and fruit nitrate content on vaso-
dilation we determined the eect of juice based diet on plasma and urine nitric oxide (NO) concentrations. Both
plasma and urine NO were signicantly increased 3-fold for plasma and 5-fold for urine at 4 days compared to
baseline (p = 1.0E−06) (Fig.4). All NO values returned to baseline concentrations at 17 days (Fig.4).
General well-being score. All participants completed a general well-being questionnaire at baseline, day 4
and at the end of study. ere was no dierence in well-being score aer the 3-day juice period (82.4 ± 14) com-
pared to baseline (82.2 ± 10). However, at the end of the study the well-being score (87 ± 11) was signicantly
increased (p = 0.006) (Fig.5).
Discussion
e scientic base for the popularity of juice based diets has not been explored. Here we demonstrated that a
3-day juice-based diet of drinking 6 bottles of fruit/vegetable juice blends (16 oz ea) resulted in a signicant
Figure 1. Eect of juice based diet on body weight (a) and body mass index (b). Data are mean ± SEM, n = 20.
One way repeat measures ANOVA was performed. Dierence to baseline is indicated; *p < 0.05.
Figure 2. Eect of juice based diet on fecal microbiota (phylum) comparing baseline (day 0), aer the
juice intervention (day 4) and aer two weeks of customary diet (day 17). Values are mean ± SEM (n = 20).
Dierence to baseline is indicated by *p < 0.05.

www.nature.com/scientificreports/
4
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
decrease in body weight (p = 2.0E−05). e observed weight loss remained signicant and persisted over the fol-
lowing 2 weeks and may be related to changes in the microbiome (p = 0.003).
e two most abundant bacterial phyla in humans are Firmicutes (40–60%) and Bacteroidetes (20–40%)10.
Relative proportional abundance of Firmicutes has been associated with increased body weight and that of
Bacteroidetes with low body weight. e 3-day juice based diet of only drinking 6 bottles of fruit/vegetable juice
blends induced signicant changes in the intestinal microbiota. e proportional abundance of the phylum
Firmicutes was signicantly decreased, while Bacteroidetes was signicantly increased when comparing end of
juice based diet (day 4) to baseline (Table1). Human Bacteroides species are able to degrade diverse plant ber and
complex polysaccharides, including pectin and xylans from fruit and vegetable16, 17. Bacteroides ovatus, B. thetaio-
taomicron and B. uniformis ferment a particularly wide range of complex polysaccharides16. In the present study
eight Bacteroides species were signicantly increased aer the juice based diet. A signicant increase in the pro-
portion of these Bacteroides species (B. ovatus, B. acidifaciens and B. xylanisolvens) was also observed in another
human intervention study in individuals with metabolic syndrome who included the prebiotic resistant starch
in their diet18. In that study the increase in Bacteroides was associated with a decrease in markers of metabolic
Genus Day 0 Day 7 Day 17
Anaeroarcus 0.2 ± 0.03b0.3 ± 0.05a0.1 ± 0.02b
Bacteroides 25.6 ± 3.4b37.0 ± 3.5a30.2 ± 2.7a,b
Barnesiella 0.4 ± 0.15b0.8 ± 0.3a0.6 ± 0.2a,b
Bilophila 0.6 ± 0.3a,b 0.5 ± 0.2b1.2 ± 0.4a
Butyricimonas 0.1 ± 0.05a,b 0.2 ± 0.07a0.1 ± 0.03b
Dialister 3.6 ± 1.19a2.4 ± 0.9b3.6 ± 1.5a,b
Eisenbergiella 0.2 ± 0.06a0.1 ± 0.01b0.2 ± 0.05a
Erysipelatoclostridium 0.2 ± 0.04a,b 0.3 ± 0.11a0.1 ± 0.02b
Faecalibacterium 6.7 ± 1.9a,b 4.2 ± 1.02b7.5 ± 1.4a
Haemophilus 0.02 ± 0.008a0.002 ± 0.001b0.007 ± 0.004a,b
Halospirulina 0.02 ± 0.007b0.2 ± 0.07a0.09 ± 0.07b
Odoribacter 0.3 ± 0.11b0.6 ± 0.1a0.4 ± 0.09b
Oscillospira 2.3 ± 1.0a2.1 ± 0.5b2.6 ± 0.84a,b
Parabacteroides 2.5 ± 0.66a2.4 ± 0.52a,b 3.3 ± 0.6b
Paraprevotella 0.2 ± 0.12b0.8 ± 0.4a0.8 ± 0.38a,b
Ralstonia 0.002 ± 0.000a0.001 + 0.000b0.001 ± 0.000a,b
Ruminiclostridium 1.0 ± 0.28a0.5 ± 0.15b0.5 ± 0.12a,b
Streptococcus 1.5 + 0.63a0.1 + 0.03b0.3 + 0.09b
Subdoligranulum 4.1 ± 1.16a1.5 ± 0.34b2.5 ± 0.7a,b
Table 2. Eect of juice based diet on fecal microbiota (genus). is table includes only signicantly altered
bacteria. Supplementary Table S2 includes bacteria that did not change signicantly. Data presents proportions
of bacteria as percent of total count. Values are mean ± SEMs (n = 20). a,bLabeled means without a common
superscript letter dier, p < 0.05.
Figure 3. Eect of juice based diet on urinary malondialdehyde concentration. Data are mean ± SEM. One way
repeat measures ANOVA was performed. e dierence to baseline is indicated **p < 0.001.

www.nature.com/scientificreports/
5
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
syndrome (cholesterol, blood pressure, etc.). In another study B. thetaiotaomicron was used as probiotic in com-
bination with a prebiotic in rats fed a high fat diet, resulting in a signicant decrease in weight and postprandial
plasma triglyceride levels19. Products of ber/complex carbohydrate metabolism are short chain fatty acids, which
have been shown to play an important role in the cardiovascular health benets of ber consumption20.
e only other human study investigating the eect of juice based diet on changes in the gut microbiota
was published by Remely et al.21. ey observed changes in the stool microbiome in participants following a
traditional program in an Austrian Monastery including a small amount of cereal, vegetable, fruit and herbal
tea21. Aer one week of the program an increase in microbial diversity was observed as well as the abundance
of Akkermansia and Bidobacteria was increased and the abundance of Enterobacteria and Lactobacilli was
decreased as determined by quantitative PCR of 16S rDNA of the bacteria of interest21.
In the present study the consumption of the juices was associated with a signicant increase in plasma and
urine nitric oxide. Nitric oxide (NO), a known vasodilator, is an important factor in cardiovascular disease.
Impaired eNOS activity and lack of bioavailable NO play a role in the development and progression of endothe-
lial dysfunction that leads to arterial stiness and an increase in blood pressure22. A study by Velmurugan et al.
Figure 4. e eect of juice based diet on (a). Plasma nitric oxide and (b). Urinary nitric oxide. Data are
mean ± SEM (n = 20). One way repeat measures ANOVA was performed. e dierence to baseline is indicated
***p < 1.0E−08.
Figure 5. Eect of juice based diet on Wellness Score. Data are mean ± SEM (n = 20). One way repeat measures
ANOVA was performed. e dierence to baseline is indicated *p < 0.05.

www.nature.com/scientificreports/
6
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
demonstrated that the consumption of dietary nitrate from beetroot juice resulted in signicant increase in ultra-
sound ow-mediated dilatation in hypercholesterolemic patients23. Dietary interventions leading to increased
blood NO concentrations may be benecial in the prevention of heart disease.
Sources of nitric oxide (NO) are either the endogenous formation through the endothelial arginine nitric
oxide synthase (eNOS) pathway or from dietary sources in form of nitrate24. One of the juices consumed during
the juice fast was prepared from green leafy vegetable and one from beetroot, which are both good dietary sources
of nitrate24. Polyphenols have been shown to enhance eNOS activity22, 24. Due to the high polyphenol content
of the fruit/vegetable juices stimulation of endogenous eNOS activity may have contributed to the increase in
NO. Endothelial dysfunction is one of the hallmark warning signs of cardiovascular disease and is involved in
disease states such as atherosclerosis and hypertension, in which the normal function of endothelium is severely
aected25. Dietary nitrate is mostly reduced to nitrite by nitrate-reducing bacteria in the oral cavity but also by
bacteria in the lower intestinal tract24, 26, 27. In the present study, however, our functional analysis of the microbial
metagenome using PICRUST showed no dierence in the nitrogen metabolism (p = 0.24).
We observed a decrease in the concentration of urinary lipid peroxidation product malondialdehyde (MDA)
aer the 3-day juice fast. is may have been based on the high polyphenol content providing antioxidant pro-
tection for lipids in the intestine during the digestion. Similar eects on postprandial lipid peroxidation were
observed by our laboratory and others utilizing other polyphenol sources. For example the addition of a spice
mix or red wine to a high fat meal dramatically decreased postprandial urinary MDA formation28, 29. Another
study by Bub et al. also supports the antioxidant eect of polyphenol rich fruit juices30. In this randomized,
crossover-design study in healthy men, a daily consumption of polyphenol-rich juices (330 ml/d) consumed for
2 weeks decreased plasma malondialdehyde compared to baseline30. In addition, urine MDA may have been
decreased based on the low fat content of the juice fast (15% calories from fat).
e vegetable/fruit juice based diet consumption showed a delayed response to improved well-being aer par-
ticipants returned to their regular diet for two weeks. Since the microbiota changes mostly returned to baseline
levels, it appears unlikely that changes in the well-being were related to the gut microbiota. As mediator of mood,
appetite and sleep, serum levels of the monoamine neurotransmitter serotonin could play a role in well-being.
However, no signicant change in serum serotonin levels were found (data not shown). Possibly participants felt
good about the weight loss that was maintained through the two weeks aer returning to the customary diet.
In summary the 3-day vegetable/fruit juice-based diet induced significant changes in the intestinal
microbiota which were associated with weight loss. Further mechanistic studies are required to confirm
that changes in the microbiota are directly linked to weight loss. The juice- based diet also, significantly
increased serum and urine NO and decreased a marker of lipid oxidation. Future studies will need to be
performed to determine if these biochemical changes are associated with vasodilation and improved cardi-
ovascular health.
Methods
Study design and juice intervention. e study was carried out in accordance with the guidelines of the
Oce for Protection of Research Subjects of the University of California, Los Angeles. e clinical protocol was
approved by the internal review board (IRB) of the University of California, Los Angeles. e study was registered
at the NIH Clinical Trial Registry: NCT02377063 on 2/18/2015.
All subjects provided written informed consent before the study began. e study was divided into three
periods: 2-week run-in period, 3-day juice fast and 2-week follow-up period. Healthy subjects who consume <3
servings of fruits/vegetables were recruited to the study. During the run-in period the participants were asked to
continue their usual diet and refrain from consuming vitamins and antibiotics. During the intervention period
all participants only consumed 6 bottles of vegetable/fruit juice daily for 3 days. e 6 dierent types of vegeta-
ble/fruit blends were prepared from the following fruits and vegetables: Green mixes were blended from apple,
cucumber, celery, romaine lettuce, lemon and limited amount (<2%) of spinach, kale and parsley. One of the
green mixes also contained ginger. e root mix was a juice blend of apple, lemon, ginger and beet. e citrus mix
contained apple, pineapple, and very limited amount (<1%) of lemon and mint. Lemon cayenne water consisted
of ltered water with cayenne and lemon and “vanilla almond (VA)” was a blend of almond, dates, sea salt and
vanilla bean. e nutritional facts are listed in Supplementary Table S1. Each bottle contained 16 oz. or 2 servings.
e total caloric content of all 6 bottles providing 1310 kcal per day. e juices provided 38 g of total fat per day
(266 kcal, 20% calories from fat) and 28 g of ber.
During the follow-up period participants resumed their usual diet consumed during the run-in period follow-
ing the same fruit and vegetable serving, vitamin and antibiotic restrictions. Juices were obtained from Pressed
Juicery (Santa Monica, CA). Nutrition facts are provided in Supplementary Table S1. On day 0, 4 and 17 partici-
pants completed a Psychological General Well Being Questionnaire (http://www.opapc.com).
Subjects. Twenty-ve healthy women and men 18–50 years of age with a custom diet including <3 servings
of fruits/vegetables per day were recruited through local advertisement. Subjects with a history of diabetes mel-
litus on medications, hyperlipidemia on medications, or other serious medical condition within 6 months prior
to screening were excluded. In addition, individuals using antibiotics or laxatives during 2 months prior to the
study were excluded.
Blood and urine collection. Fasting. EDTA blood samples and 24 hr urine samples were collected at
baseline, day 4 and day 17. Blood was centrifuged 1500 × g for 10 min at 10 °C and plasma stored at 80 °C until
analysis.

www.nature.com/scientificreports/
7
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
Stool Collection. A total of three stool samples were collected from each subject: at baseline, day 4 and day
17. Each time an aliquot of the stool specimen was collected by the participant and delivered to the UCLA Center
for Human Nutrition in a cooler within a few hours of collection. At the laboratory stool was aliquoted into
smaller vials and frozen immediately and stored at −80 °C.
Bacterial DNA Sequencing. DNA from stool was extracted using the MoBio power soil DNA isolation kit
(MoBio Laboratories, Inc., Carlsbad, CA). e quality and quantity of the DNA was conrmed using a Nanodrop
1000 (ermo Fisher Scientic, Wilmington, DE). e 16S rRNA gene V4 variable region PCR primers 530/926
with barcode on the forward primer were used in a 30 cycle PCR using the HotStarTaq Plus Master Mix Kit
(Qiagen, USA) under the following conditions: 94 °C for 3 min, followed by 28 cycles of 94 °C for 30 s, 53 °C for
40 s and 72 °C for 1 min, aer which a nal elongation step at 72 °C for 5 min was performed. Aer amplication,
PCR products are checked in 2% agarose gel to determine the success of amplication and the relative inten-
sity of bands. Sequencing was performed at MR DNA (www.mrdnalab.com, Shallowater, TX, USA) on a MiSeq
(Illumina, San Diego, CA) following the manufacturer’s guidelines. Sequence data were processed using a pro-
prietary analysis pipeline (MR DNA, Shallowater, TX, USA). Operational taxonomic units (OTUs) were dened
by clustering at 3% divergence (97% similarity). Final OTUs were taxonomically classied using BLASTn against
a curated GreenGenes database31. Within community diversity (α-diversity) was calculated using Quantitative
Insights Into Microbial Ecology (QIIME) software package32. Analysis of α-diversity (Shannon index) was
performed by a one-way ANOVA. β-diversity was measured by calculating the weighted UniFrac distances33
using QIIME default scripts, and weighted UniFrac PCoA biplot was visualized using EMPeror. Statistical dif-
ference between dierent time points was analyzed by PERMDISP. Microbiome data was further analyzed using
PICRUST (phylogenetic investigation of communities by reconstruction of unobserved states) to predict the
functional composition of the microbial community’s metagenome from its 16S prole by using marker gene data
and a database of reference genomes34.
Body Composition. Body composition was measured using the Tanita-BC418 bioelectrical impedance ana-
lyzer (Tanita Corp., Japan).
Malondialdehyde. Urine malondialdehyde concentration was determined in form of thiobarbituric
acid (TBA) reactive substances by deproteinization and derivatization with TBA according to the method of
Korchazhkina et al.35. In brief, 0.6 mL urine was mixed with 0.4 mL of water and 3 mL of H3PO4 (1%V/V) and
vortexed. One ml of TBA solution was added and incubated in a 100 °C water bath for 60 min. Samples were
cooled on ice and centrifuged at 18,000 × g for 10 min. An aliquot of 50 µL supernatant was injected into the
HPLC system for analysis. Chromatographic determinations were performed on a Waters 2690 HPLC (Waters,
Milford, MA) equipped with a Waters 474 uorescence detector (Waters, Milford, MA) at Ex (ex-citation) 515 nm
and Em (emission) 550. An Agilent Zorbax SB C18 reversed-phase column (150 mm, 4.6-mm inside diameter;
3.5-µm particles; Waters, Milford, MA) was used for separation at an ambient temperature. A linear gradient from
10% A and 90% B to 60% A and 40% B over 20 min (A: acetonitrile and B: 0.1% phosphoric acid in water) at a
ow rate of 0.8 mL/min was used for the elution. e reagent 1,1,3,3-tetramethoxypropane was used to prepare
a standard curve.
Nitric oxide. Nitric oxide (NO) is oxidized in the body to the metabolites nitrate (NO3−) and nitrite (NO2−),
which are excreted in the urine. Nitrate/nitrite concentrations were measured in the form of NO in urine and
serum samples by an ozone chemiluminescence method (Sievers Instruments, Boulder, CO). Serum samples
were mixed with ethanol in a ratio of 1:2 and stored at 4 °C overnight. Aer centrifugation for 20 min at 17 000 × g
an aliquot of the supernatant was injected into the reaction chamber of the NO analyzer. Urine was centrifuged
at the same speed and injected. Total nitrate and nitrite were analyzed by injecting 5 μL of each microdialysate
sample into a purge vessel containing a solution of vanadium (III) chloride (50 mmol/L) in hydrochloric acid
(1 mol/L) at 95 °C, continuously purged with a stream of nitrogen gas, connected to a Sievers 280i Nitric Oxide
Analyser (GE Analytical Instruments, Boulder, CO, USA). e concentration was determined in comparison to
a sodium nitrate standard calibration curve.
Urine creatinine. e urinary creatinine concentration was measured by using a Stanbio Direct Creatinine
LiquiColor Kit (Stanbio LaboratoryBoerne, TX, USA) according to manufacturer’s instructions. A creatinine
standard calibration curve was generated by plotting the absorbance measured at 510 nm of a series concentration
of creatinine standard solutions reacting with working reagent provided by the kit. e urinary creatinine level
was measured in the same manner aer a proper dilution with distilled water.
Statistics. Statistical analyses were performed using IBM SPSS Statistics version 22 (IBM Corporation,
Armonk, NY, USA). Data were evaluated with One-way repeated measures ANOVA with a Bonferroni’s post-test
if the assumption are met, and those without were analyzed using non-parametric test (Friedman test with a
Dunn’s post-test). P-values < 0.05 were considered statistically signicant. Correlation was evaluated using
GraphPad Prism 6 (La Jolla, CA).
References
1. Horne, B. D., Muhlestein, J. B. & Anderson, J. L. Health eects of intermittent fasting: hormesis or harm? A systematic review. e
American journal of clinical nutrition 102, 464–470, doi:10.3945/ajcn.115.109553 (2015).
2. Tome-Carneiro, J. & Visioli, F. Polyphenol-based nutraceuticals for the prevention and treatment of cardiovascular disease: eview
of human evidence. Phytomedicine: international journal of phytotherapy and phytopharmacology, doi:10.1016/j.phymed.2015.10.018
(2015).

www.nature.com/scientificreports/
8
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
3. Abuajah, C. I., Ogbonna, A. C. & Osuji, C. M. Functional components and medicinal properties of food: a review. Journal of food
science and technology 52, 2522–2529, doi:10.1007/s13197-014-1396-5 (2015).
4. Li, A. N. et al. esources and biological activities of natural polyphenols. Nutrients 6, 6020–6047, doi:10.3390/nu6126020 (2014).
5. Dahl, W. J. et al. Health Benets of Fiber Fermentation. Journal of the American College of Nutrition 1–10, doi:10.1080/07315724.20
16.1188737 (2017).
6. Simpson, H. L. & Campbell, B. J. eview article: dietary bre-microbiota interactions. Alimentary pharmacology & therapeutics 42,
158–179, doi:10.1111/apt.13248 (2015).
7. orpela, . et al. Gut microbiota signatures predict host and microbiota responses to dietary interventions in obese individuals. PloS
one 9, e90702, doi:10.1371/journal.pone.0090702 (2014).
8. Tilg, H. & Moschen, A. . Microbiota and diabetes: an evolving relationship. Gut 63, 1513–1521, doi:10.1136/gutjnl-2014-306928
(2014).
9. Tuohy, . M., Fava, F. & Viola, . ‘e way to a man’s heart is through his gut microbiota’–dietary pro- and prebiotics for the
management of cardiovascular ris. e Proceedings of the Nutrition Society 73, 172–185, doi:10.1017/S0029665113003911
(2014).
10. Million, M., Lagier, J. C., Yahav, D. & Paul, M. Gut bacterial microbiota and obesity. Clinical microbiology and infection: the ocial
publication of the European Society of Clinical Microbiology and Infectious Diseases 19, 305–313, doi:10.1111/1469-0691.12172
(2013).
11. Etxeberria, U. et al. eshaping faecal gut microbiota composition by the intae of trans-resveratrol and quercetin in high-fat sucrose
diet-fed rats. e Journal of nutritional biochemistry 26, 651–660, doi:10.1016/j.jnutbio.2015.01.002 (2015).
12. oopchand, D. E. et al. Dietary Polyphenols Promote Growth of the Gut Bacterium Aermansia muciniphila and Attenuate High-
Fat Diet-Induced Metabolic Syndrome. Diabetes 64, 2847–2858, doi:10.2337/db14-1916 (2015).
13. Duenas, M. et al. A survey of modulation of gut microbiota by dietary polyphenols. BioMed research international 2015, 850902–15,
doi:10.1155/2015/850902 (2015).
14. Turnbaugh, P. J. et al. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 444, 1027–1031,
doi:10.1038/nature05414 (2006).
15. Stenman, L. ., Burcelin, . & Lahtinen, S. Establishing a causal lin between gut microbes, body weight gain and glucose
metabolism in humans - towards treatment with probiotics. Benecial microbes 1–12, doi:10.3920/BM2015.0069 (2015).
16. Flint, H. J., Scott, . P., Duncan, S. H., Louis, P. & Forano, E. Microbial degradation of complex carbohydrates in the gut. Gut
microbes 3, 289–306, doi:10.4161/gmic.19897 (2012).
17. avcheev, D. A., Godzi, A., Osterman, A. L. & odionov, D. A. Polysaccharides utilization in human gut bacterium Bacteroides
thetaiotaomicron: comparative genomics reconstruction of metabolic and regulatory networs. BMC genomics 14, 873,
doi:10.1186/1471-2164-14-873 (2013).
18. Upadhyaya, B. et al. Impact of dietary resistant starch type 4 on human gut microbiota and immunometabolic functions. Scientic
reports 6, 28797, doi:10.1038/srep28797 (2016).
19. Olli, . et al. Independent and Combined Eects of Lactitol, Polydextrose, and Bacteroides thetaiotaomicron on Postprandial
Metabolism and Body Weight in ats Fed a High-Fat Diet. Frontiers in nutrition 3, 15, doi:10.3389/fnut.2016.00015 (2016).
20. Marques, F. Z. et al. High Fibre Diet and Acetate Supplementation Change the Gut Microbiota and Prevent the Development of
Hypertension and Heart Failure in DOCA-Salt Hypertensive Mice. Circulation, doi:10.1161/CICULATIONAHA.116.024545
(2016).
21. emely, M. et al. Increased gut microbiota diversity and abundance of Faecalibacterium prausnitzii and Aermansia aer fasting:
a pilot study. Wiener linische Wochenschri 127, 394–398, doi:10.1007/s00508-015-0755-1 (2015).
22. Cutler, B. ., Petersen, C. & Anandh Babu, P. V. Mechanistic insights into the vascular eects of blueberries: Evidence from recent
studies. Molecular nutrition & food research, doi:10.1002/mnfr.201600271 (2016).
23. Velmurugan, S. et al. Dietary nitrate improves vascular function in patients with hypercholesterolemia: a randomized, double-blind,
placebo-controlled study. e American journal of clinical nutrition 103, 25–38, doi:10.3945/ajcn.115.116244 (2016).
2 4. Bondonno, C. P., Cro, . D., Ward, N., Considine, M. J. & Hodgson, J. M. Dietary avonoids and nitrate: eects on nitric oxide and
vascular function. Nutrition reviews 73, 216–235, doi:10.1093/nutrit/nuu014 (2015).
25. Chistia ov, D. A., Orehov, A. N. & Bobryshev, Y. V. Endothelial Barrier and Its Abnormalities in Cardiovascular Disease. Frontiers
in physiology 6, 365, doi:10.3389/fphys.2015.00365 (2015).
26. Weitzberg, E. & Lundberg, J. O. Novel aspects of dietary nitrate and human health. Annual review of nutrition 33, 129–159,
doi:10.1146/annurev-nutr-071812-161159 (2013).
27. Tiso, M. & Schechter, A. N. Nitrate reduction to nitrite, nitric oxide and ammonia by gut bacteria under physiological conditions.
PloS one 10, e0119712, doi:10.1371/journal.pone.0119712 (2015).
28. G oreli, S., Ligumsy, M., ohen, . & anner, J. A novel function of red wine polyphenols in humans: prevention of absorption of
cytotoxic lipid peroxidation products. FASEB journal: ocial publication of the Federation of American Societies for Experimental
Biology 22, 41–46, doi:10.1096/.07-9041com (2008).
29. Li, Z. et al. Antioxidant-rich spice added to hamburger meat during cooing results in reduced meat, plasma, and urine
malondialdehyde concentrations. The American journal of clinical nutrition 91, 1180–1184, doi:10.3945/ajcn.2009.28526
(2010).
30. Bub, A. et al. Fruit juice consumption modulates antioxidative status, immune status and DNA damage. e Journal of nutritional
biochemistry 14, 90–98, doi:10.1016/S0955-2863(02)00255-3 (2003).
31. DeSantis, T. Z. et al. Greengenes, a chimera-checed 16S rNA gene database and worbench compatible with AB. Applied and
environmental microbiology 72, 5069–5072, doi:10.1128/AEM.03006-05 (2006).
32. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nature methods 7, 335–336,
doi:10.1038/nmeth.f.303 (2010).
33. Lozupone, C. & night, . UniFrac: a new phylogenetic method for comparing microbial communities. Applied and environmental
microbiology 71, 8228–8235, doi:10.1128/AEM.71.12.8228-8235.2005 (2005).
34. Langille, M. G. et al. Predictive functional proling of microbial communities using 16S rNA marer gene sequences. Nature
biotechnology 31, 814–821, doi:10.1038/nbt.2676 (2013).
35. orchazhina, O., Exley, C. & Andrew Spencer, S. Measurement by reversed-phase high-performance liquid chromatography of
malondialdehyde in normal human urine following derivatisation with 2,4-dinitrophenylhydrazine. Journal of chromatography.
B, Analytical technologies in the biomedical and life sciences 794, 353–362, doi:10.1016/S1570-0232(03)00495-1 (2003).
Acknowledgements
Supported by departmental funds from the Center for Human Nutrition, Department of Medicine, David Geen
School of Medicine, University of California, Los Angeles.

www.nature.com/scientificreports/
9
Scientific RepoRts | 7: 2167 | DOI:10.1038/s41598-017-02200-6
Author Contributions
S.M.H. and Z.L. wrote the main manuscript text, J.Y. and A.L. performed QIIME analysis, R.P.L. and J.H.
performed chemical analyses, P.S., M.H. and G.T. coordinated clinical study, Q.Y.L. evaluated juices, D.H.
participated in manuscript preparation. All authors reviewed the manuscript.
Additional Information
Supplementary information accompanies this paper at doi:10.1038/s41598-017-02200-6
Competing Interests: e authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2017