Starch/Stärke 2013, 65, 509–516 DOI 10.1002/star.

201200166 509

RESEARCH ARTICLE

Effect of resistant starch structure on short-chain

fatty acids production by human gut microbiota
fermentation in vitro
Zhongkai Zhou1, Xiaohong Cao1 and Julia Y. H. Zhou2
1
School of Food Engineering and Biotechnology, Tianjin University of Science and Technology, Tianjin, P. R. China
2
School of Pharmacy and Medical Science, University of South Australia, Adelaide, Australia

Resistant starch (RS) from uncooked and cooked high amylose starch (HAS) was used to Received: August 4, 2012
investigate the effect of RS structure on the production of short-chain fatty acids (SCFA) using Revised: September 24, 2012
a static anaerobic in vitro system. The pH value of culture correlated well to the production of Accepted: September 27, 2012
total SCFA after the fermentation (R2 ¼ 0.969). Most importantly, fermentation of RS from
thermally treated starch under no moisture condition produced the highest concentration of
SCFA and the greatest ratios of butyrate/acetate and butyrate/total SCFA, followed by the
fermentation of RS from uncooked HAS. FTIR analysis suggests that all the RS exhibited a
relatively higher organized structure compared to its corresponding original starch. The
analysis of molecular structure by HPLC showed that the most pronounced difference in
molecular composition of RS between the group favoring greater butyrate production and
the group supporting a lower butyrate production was that larger molecules (amylopectin
fraction) still existed in the former group but not in the latter group. Thus, it can be concluded
that RS structure, in particular molecular structure is one of the key factors manipulating
SCFA production in amount and proportion.

Keywords:
Butyrate / Gut microbiota fermentation / Molecular structure / RS / Short-chain fatty acids

1 Introduction RS as a substrate for stimulating colonic fermentation

and promoting production of short-chain fatty acids
RS is defined as the ‘‘total amount of starch, and the (SCFA) [2–5]. In terms of production, acetic, propionic,
products of starch degradation that resists digestion in and butyric are the major SCFA produced in the human
the small intestine of healthy people’’ [1]. In the large colon. Physiological studies found that SCFA can modu-
intestine RS acts as a substrate for microbial fermentation late colonic muscular activity and stimulate blood flow
to produce a number of metabolites that are considered to the colon, and they also appear to lower the risk
beneficial for the host. In recent years, an increasing of pathogen overgrowth [6, 7]. Of the principal SCFA,
number of studies have focused on the importance of butyrate is thought to be pivotal for human colonic
health, as it has been shown to promote growth of normal
colonocytes and enhance apoptosis in colorectal cancer
Correspondence: Dr. Zhongkai Zhou, School of Food Engineering cells in vitro [8–11].
and Biotechnology, Tianjin University of Science and Technology,
There is some evidence that RS, relative to other poten-
Tianjin 300457, P. R. China
E-mail: zhongkai_zhou@[Link] tial substrates such as non-starch polysaccharides, favors
Fax: þ86-22-60601371 increasing production of butyrate [12–15]. However, this
may not be a characteristic of all RSs [7]. For instance,
Abbreviations: HAS, high amylose starch; HTT, hydrothermal starches from various botanical sources were fermented
treatment; NTT, non-moisture thermal treatment; SCFA, short-
chain fatty acid; SE-HPLC, size-exclusion high performance liquid differently by the microbiota of the large intestine, produc-
chromatography ing gas and SCFA in different proportions and at different

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

510 Z. Zhou et al. Starch/Stärke 2013, 65, 509–516

rates depending on their origin [16]. These differences properties) and its physiological functionality (e.g. SCFA
may be due in part to the experimental method used, production). In this paper, high amylose starch (HAS) in
sample preparation, type, amount, and structure of RS, uncooked (native) and cooked forms are used to study the
and on the duration of feeding. Nevertheless, under tightly influence of starch structure on the production of SCFA
controlled experimental conditions it seems that the nature using a static anaerobic in vitro system inoculated with a
of the starch plays a very important role in modulating fresh fecal slurry to simulate large bowel fermentation in
bacterial fermentation, especially concerning the amount humans.
and type of SCFA that are produced [17].
Quantitative data on the relationship between dietary
starch and the metabolic activity of the colonic microbiota 2 Materials and methods
are still very limited due largely to the inaccessability of
intestinal contents of humans. Much information on the 2.1 Starch sample
role of RS as a fermentative substrate derives from studies
that have involved the use of human fecal inocula to fer- A high amylose maize starch (HAS, Hi-maizeTM, National
ment indigestible polysaccharides in vitro as a method of Starch and Chemical Company, NSW, Australia) was
quantifying the production of SCFA and lactate [5, 15, 18, used in this study. High amylose starch varieties (e.g.
19]. A major deficiency is lack of knowledge of the relation- HAS) usually associated have a higher RS content
ship between starch/RS structure (both in physical and than conventional starch varieties which are significantly
chemical properties), SCFA production and its compo- degraded during digestion in the human gut, and a
sition, and the large bowel microbiota profile. This is crucial very limited fraction can reach the colon. Thus, it was
because the effects of RS appear to be mediated largely only possible to obtain sufficient amount of RS from
through fermentation products rather than physical bulk- HAS and its residual was processed for physical and
ing. In this study, starch chemical and physical properties chemical analyses.
were investigated using size-exclusion high performance
liquid chromatography (SE-HPLC) and FTIR. Previous 2.2 Starch treatment/sample preparation
studies [20, 21] indicated that some of the bonds in
FTIR spectrum were sensitive to changes in the degree The preparations of non-moisture thermal treatment (NTT)
of arrangements at molecular level. Ordered and amor- starch and hydrothermal treatment (HTT) starch were
phous structures of starch are represented at 1047 and developed in this study. The methodologies for sample
1022 cm1 bands of the infrared spectrum. In order to preparations were briefly described in Table 1. Under
express the relative organization of starch structure, the different thermal treatment conditions, four groups of
ratio of absorbance at 1047 and 1022 cm1 was used. starches were obtained (Table 1). Prior to all analyses,
Hence, these could be used as a tool to study starch all samples were freeze-dried after preparation and were
structural organization. ground using a Cyclone Sample Mill (UDY Corporation,
Thus, the objective of this study was to investigate the Fort Collins, CO) and passed through a 0.5 mm sieve
relationship between RS structure (physical and chemical screen.

Table 1. Sample preparation: starch thermal treatment under different conditions

Sample group Sample preparation Sample no.

Group 1 Native (uncooked). Native (uncooked). 1
Group 2 Thermal treatment under no moisture Starch was thermally treated in an oven 2
condition (NTT)a). (at 1358C for 3.5 h).
Group 3 Hydrothermal treatment at a lower Starch was hydrothermally treated in an autoclave 3
moisture content (25%) (HTT)b). (at 1218C for 20 min) once.
Group 4 Hydrothermal treatment at a higher Starch was hydrothermally treated in an autoclave 4
moisture content (60%) (HTT). (at 1218C for 20 min) once.
Starch was hydrothermally treated in an autoclave 5
(at 1218C for 20 min) for 2 times.
Starch was hydrothermally treated in an autoclave 6
(at 1218C for 20 min) for 3 times.

a) NTT: thermal treatment under no moisture conditions.

b) HTT: thermal treatment under moisture conditions.

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

Starch/Stärke 2013, 65, 509–516 511

2.3 Starch digestion and RS preparation AG1 501-X8, USA) and incubating at 408C for 2 h with
occasional shaking. After centrifuged at 10 000 rpm for
Two hundred milligrams of starch material from uncooked 10 min, the clear supernatant was collected for HPLC
and cooked forms (see Table 1) was pre-digested using an analysis. The HPLC system comprised a GBC pump
in vitro system developed by CSIRO (Human Nutrition, (LC 1150, GBC Instruments, Vic, Australia) equipped
Adelaide, Australia) to simulate human small intestinal with Auto Sampler (GBC, LC1610) and Evaporative
starch digestion. Briefly, samples were mixed with artificial Light Scattering (ELS) Detector (ALLTech, USA). The
saliva and the resultant bolus incubated with pancreatic UltrahydrogelTM 250 column and guard column (Waters,
and gastric enzymes at physiological pH and temperature. 7.8 mm  300 mm, made in Japan) were maintained
After incubation, the supernatant was removed by centri- at 358C. Ammonium acetate buffer (0.05 M; pH 5.2)
fuge and the resultant RS was recovered and used as was used as mobile phase at a flow rate of 0.6 mL/min.
substrate for fermentation. The amount of residual starch A 40 mL-aliquot of the supernatant was used for
in the digesta (RS) was determined using Megazyme injection. Conditions for ELS detector operation were: tube
assay kit (GOPOD method, manufacture instruction). temperature: 1158C; N2 gas flow rate: 2.0 L/min; gain: 16;
After the digestion, around 90 mg of RS substrate from impactor: on.
each treated HAS was collected and added to the
fermentors. 2.6 In vitro fermentation by human gut bacteria

2.4 FTIR for analysis of starch structural A 24 h in vitro anaerobic static batch fermentation system
organization was used to investigate SCFA production of each RS.
Fresh human fecal material was used as inoculum.
FTIR spectrum of each starch sample was recorded on a Samples were collected from two separate individuals,
Varian spectrometer (Model: Excalibur 3100) equipped both of whom were apparently healthy with no history of
with a cooled deuterated triglycine sulfate (DTGS) antibiotic treatment over the preceding 3-month period,
detector. The measurement was performed on a consuming their usual diet. Feces were processed soon
MIRacleTM attenuated total reflectance (ATR) crystal after defecation and a 10% w/v inoculum prepared. Briefly,
plate with Digital Readout High Pressure Clamp (Pike equivalent amounts of homogenized stool from each donor
Technologies, USA). Sample was directly loaded on the were combined, mixed thoroughly, and then 150 mL of pre-
plate and scanned in the range of 3600–600 cm1 at a reduced phosphate buffer added. The slurry was mixed
resolution of 4 cm1. Prior to recording, the spectra were and the resultant suspension constantly stirred as aliquots
transformed against an empty cell as background. Spectra of the homogenate were removed and dispensed into
were ATR deconvoluted and baseline corrected using individual pre-sterilized McCartney vials.
Varian Resolutions Pro software. Quadruplicate incubations were set up for the blank and
each test. In the case of the blank, no additional substrate
2.5 SE-HPLC for analysis of RS molecular was added to the vials. Substrates were hydrated for 1 h
composition at 48C in 9 mL of fermentation media. Fermentation
medium contained the following (per liter of distilled water):
A 10-mL aliquot of aqueous dimethylsulfoxide (90:10, 2.5 g trypticase, 125 mL micromineral solution (132 g
DMSO/water, v/v) was added to 50 mg of freeze-dried CaCl2  2H2O, 100 g MnCl2  4H2O, 10 g CoCl2  6H2O
RS in 25 mL test tube. Each tube was capped and placed and 80 g FeCl3  6H2O per liter distilled water.), 250 mL
in a boiling water bath for 60 min and then cooled to room buffer solution (4 g (NH4)HCO3 and 35 g NaHCO3 per liter
temperature before centrifuging at 2095  g for 15 min. distilled water), 250 mL macromineral solution (5.7 g
The supernatant was collected into a separate 75 mL Na2HPO4 and 0.6 g MgSO4  7H2O per liter of distilled
polypropylene centrifuge tube and the dissolved starch water) and 1.25 mL resazurine solution 0.1% w/v. To 1L
precipitated by adding 30 mL of 95% ethanol. The tube of fermentation medium 33.5 mL of reducing solution
was then centrifuged (2095  g for 15 min), the super- (6.25 g cysteine hydrochloride, 6.25 g Na2S  9H2O and
natant discarded and the starch precipitate stored for 40 mL NaOH 1 M per liter of distilled water), was added
analysis. and the resulting solution sterilized at 1218C for 15 min.
The RS precipitate (around 30 mg), prepared as The pH of fermentation media was adjusted to 7.2.
described above, was redissolved in 0.6 mL of 0.2 M Each fermentation vial was inoculated with 1 mL of
sodium hydroxide solution and mixed vigorously for fecal slurry containing 1  1010 viable bacteria/mL to
approximately 30 s. The solution was neutralized by achieve a final concentration of 1  109 viable bacteria
addition of sodium acetate buffer (0.6 mL; 0.05 M, pH per ferment. Each vial was then sparged with nitrogen gas
4.0) before adding ion-exchange resin (0.25 g, BioRad and capped then sealed with paraffin as an additional

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

512 Z. Zhou et al. Starch/Stärke 2013, 65, 509–516

precaution to ensure no leakage. Vials were incubated 3 Results and discussion

in a hybridization oven (XTRON HI 2002, Bartelt
Instruments) at 378C and rocked gently (19 rpm) over a 3.1 IR spectrum of RS made from cooked and
24 h period, after which they were frozen immediately and uncooked starches
stored. Samples were defrosted and processed; sub-
samples from each vial were taken for total and individual FTIR ratio of 1047/1022 cm1 for each starch sample was
SCFAs, DL-lactate and pH. list in Table 2. Previous studies suggested that FTIR spec-
trum was sensitive to the changes of starch conformation
2.7 SCFA and lactate analyses during various thermal treatments, specifically in the
range of 1300–900 cm1 [20, 21]. The band 1047 cm1
Two milliliters of ferment was centrifuged at 2851  g is composed of two overlapping bands, positioned at 1040
for 15 min. One milliliter of suspension was mixed and 1053 cm1. During starch molecule’s re-association,
with 50 mL of 0.1 M 2-ethylbutyric acids as internal the 1040 cm1 band appears very quickly; however the
standard and extracted by the addition of 500 mL of 1053 cm1 band requires a longer time to form. The inten-
concentrated HCl and 2 mL of diethyl ether solution. sity of the 1047 cm1 band increases as crystallinity of
Samples were vortexed for 1 min and the layers starch increases. As an example, the higher the crystallinity
separated by centrifuging for 10 min at 2095  g. The of starch the greater the intensity of the 1047 cm1
ether layer was then transferred to a clean tube and band, whereas the intensity of the 1022 cm1 band is
capped. The extraction was repeated by adding another characterized by reduction of crysallinity of starch. Thus,
1 mL of diethyl ether to the initial tube, vortexing for Absorbance at 1047 and 1022 cm1 was associated with
1 min followed by centrifuging at 2095  g for 10 min. the ordered and amorphous structures of the starch,
The second ether layer was then removed and combined respectively [20–22]. The ratio of absorbance at these
with the first. Hundred microliters N-methyl-N-t-butyldi- particular wavelengths (1047 and 1022 cm1) might be
methylsilyltrifluoroacetamide (MTBSTFA) was added to used to express the relative organization of starch structure.
800 mL of the combined ether extracts. The reaction In this study, it was found that starch molecules treated
mixture was heated to 808C for 20 min to form t-butyldi- under a higher moisture content more likely tend to reor-
methylsilyl (TBDMS) derivative, and left at room tempera- ganize and then form a more organized structure than
ture for a further 24 h to ensure complete derivatization of starch treated under a lower moisture content. Compared
lactic acid. to uncooked status, there was a slight increase in starch
Derivatives were analyzed using a gas chromatograph structural organization after three times HTTs (Fig. 1).
(6890 Agilent) fitted with a flame ionization detector, split Meanwhile, increasing HTT cycle can also enhance starch
injector and a Alltech AT-1 30 m  0.25 mm capillary col- structural organization at a higher moisture condition (i.e.
umn with 0.1 mm film thickness. Injector and detector sample no. 6>sample no. 5>sample no. 4; Table 2).
temperatures were 2758C with the column temperature
programmed from 638C for 3 min to 1908C at 108C/min. Table 2. Change in FTIR Ratio of absorbance at 1047/
Helium was the carrier gas (head pressure 100 kPa). 1022 cm1 before and after digestions
0.2 mL injections were made in the split mode (50:1 split).
The standard solution contained (mmol/L): acetate, 30; FTIR ratio: 1047/1022 cm1)b)
propionate, 20; isobutyrate, 5; n-butyrate, 20; isovalerate,
Substrate Sample Before digestion After digestion
5; n-valerate, 5; lactate, 10. The internal standard was
groupa) no. (i.e. original starch) (i.e. RS)
2-ethylbutyric acid (100 mmol/L).
Group 1 1 0.844  0.03 1.25  0.03
2.8 Statistical analysis Group 2 2 0.815  0.04 1.41  0.03
Group 3 3 0.755  0.0 1.91  0.07
RS is quantified as the difference between total starch Group 4 4 0.769  0.03 1.97  0.06
5 0.833  0.04 2.03  0.06
and readily digestible starch after digestion. All of the
6 0.855  0.0 2.01  0.07
sample analyses were performed and expressed as
means  SEM of 3 replicates. Experimental data were
a) Group 1: uncooked/native HAS; group 2: HAS was
subjected to analysis of variance using Genstat 5
thermally treated in an oven; group 3: HAS was hydro-
(release 4.1). Differences in fermentation properties of thermally treated in an autoclave at 25% moisture con-
RS from uncooked and cooked starches were assessed dition; group 4: HAS was hydrothermally treated in an
using the least significant differences (LSD) of means autoclave at 60% moisture condition.
(5% level). The difference was considered as significant b) Data expressed as mean  SEM of triplicate determi-
at p < 0.05. nations for each sample.

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

Starch/Stärke 2013, 65, 509–516 513

0.5 3.2 Molecular composition of RS

Cooked HAS
0.4 Native HAS The difference in molecular composition among the RS
substrates was examined and their chromatographs
0.3
were presented in Fig. 2a–c. The pronounced feature of
Absorbance

the structural profile among the substrates was the differ-

0.2
ence in the ratio of amylopectin/amylose. For instance, RS
from uncooked HAS had the highest amylopectin fraction
0.1
in its molecular profile, followed by RS from NTT sample.
However, for molecular profile of RS from HTT samples
0.0
(both group 3 and group 4), their amylopectin fraction was
completely absent, irrespective of moisture content during
-0.1
1100 1080 1060 1040 1020 1000 980 960 940 HTT.
Wavenumber (cm-1) Figure 2 indicated that thermal treatment altered starch
structure from exterior to interior, subsequently influencing
Figure 1. Difference in FTIR spectra between native/ RS molecular structure. The results also implied thermal
uncooked HAS (sample no. 1) and cooked HAS (sample treatment on starch, either with or without moisture, greatly
no. 6).
enhanced enzymatic susceptibility for both amylose and
amylopectin fractions, in particular for amylopectin frac-
tion. The obtained RS with different molecular structure is
Data in Table 2 demonstrated that all RS had a higher useful to study the effect of molecular structure on its
ratio of 1047/1022 cm1 compared to the starch before physiological property.
digestion (i.e. ranging from 1.25 to 2.03 after digestion
versus ranging from 0.755 to 0.855 before digestion). 3.3 pH of ferment culture and total SCFA
Because the IR ratio for RS exceeds 1 suggesting that production
every RS exhibited a relatively higher organized structure
due to the removal of the digestible starch fraction by In vitro systems have been used as an alternative to
digestion [23]. RS from thermal treated starch, irrespective invasive techniques to investigate RS fermentation.
of moisture content was of greater structural organization Static 24 h batch fermentations were set up as
than RS from uncooked (Table 2). described previously using substrates consisting of

Figure 2. Difference in molecular composition of RS from uncooked HAS (the curve on the bottom) and RS from thermally
treated starch (the curve on the top) (a: uncooked HAS versus NTT sample; b: uncooked HAS versus HTT sample at 25%
moisture condition (sample no. 3); c: uncooked HAS versus HTT sample at 60% moisture condition (sample no. 5). NTT:
thermally treated under no moisture condition; HTT: thermally treated under moisture condition.

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

514 Z. Zhou et al. Starch/Stärke 2013, 65, 509–516

RS from different treatment of HAS. Quadruplicate

fermentations were inoculated with a human fecal
microbiota suspension and incubated for 24 h under
anaerobic conditions at 378C (as described in materials
and method).
The culture medium had a pH of 7.2 at time zero.
After fermentation, all the cultures showed a decrease
in pH value (Table 3), but the magnitude of the
reduction in pH value was markedly different depending
on the RS resources. Generally, the fall in pH value
observed for fermentation of RS from NTT sample was
greater than that of RS from HTT sample. This difference
might be due to the higher concentration of net total
SCFA, in particular acetate produced by fermentation,
because acetate is the major SCFA in each ferment Figure 3. Relationship between culture pH value and net
and it has a lower pKa (4.74) than other SCFA in cultures. total SCFA concentration after gut bacterial fermentation of
each RS sample.
The pH value of batch culture correlated very well to
the production of net total SCFA after the fermentation
(R2 ¼ 0.969; Fig. 3). Acidification of the large bowel
lumen is believed to be advantageous for the host in 3.4 Principal SCFA concentration and
that it is less favorable for the growth of intestinal patho- proportion
gens and may also reduce the production and activity of
potential toxins [24]. The principal SCFAs in all instances were acetic, propionic
The net total SCFA production, estimated as mmol/L and butyric (Table 3). Other SCFAs were also detected in
of each fermented culture, is summarized in Table 3. cultures, including iso-butyric, iso-valeric, and valeric
The effect of RS resources on SCFA production acids, but in much lower concentrations and were not
showed a similar pattern to that of pH. Briefly, RS reported in this study. Of the principal SCFA, acetic acid
from NTT sample produced the highest SCFA was present at the highest concentration in all ferments.
concentration, followed by RS from uncooked HAS, Although no significant difference in acetate production
and then RS from HTT samples. Increase of thermal was found between the fermentation of RS from uncooked
treatment cycles under higher moisture content HAS and RS from NTT sample, fermentation of RS from
(group 4 samples) could significantly improve SCFA uncooked HAS produced the highest acetate concen-
production (Table 3; p < 0.001). Results obtained in tration in all instances.
this study suggested that RS from either NTT sample Fermentation of RS from NTT sample produced the
or uncooked HAS was the best ferment substrate for highest concentration of butyrate, followed by the fermen-
SCFA production. tation of RS from uncooked HAS, whereas HTT groups

Table 3. Culture pH value and net total SCFA and lactate concentrations after 24 h fermentation

SCFA (mmol/L)b) Butyric/ Butyric/

Sample Acetic total
Group no.a) pH Acetic Propionic Butyric Lactic Total (100) SCFA (%)

Group 1 1 5.55  0.3 23.44  0.63 4.34  0.08 8.15  0.42 Nil 35.93  0.35 34.77 22.68
Group 2 2 5.26  0.1 20.66  2.38 9.40  1.07 14.55  1.88 Nil 44.58  3.56 70.43 32.64
Group 3 3 5.99  0.1 18.93  1.63 2.83  0.30 3.38  0.45 0.02  0.03 25.17  1.39 17.86 13.43
Group 4 4 6.27  0.4 12.43  5.13 3.74  0.41 2.20  1.62 Nil 18.32  6.80 17.70 12.01
5 6.00  0.3 15.79  2.52 3.65  0.16 3.27  0.72 Nil 22.66  3.28 20.71 14.43
6 5.90  0 21.47  1.11 4.03  0.29 4.95  0.38 Nil 30.41  1.11 23.06 16.28

a) Sample 1: native HAS; sample 2: HAS was thermally treated in an oven; sample 3: HAS was hydrothermally treated at
25% moisture condition; sample 4: HAS was hydrothermally treated at 60% moisture condition once; sample 5: HAS was
hydrothermally treated at 60% moisture condition twice; sample 6: HAS was hydrothermally treated at 60% moisture
condition for three times.
b) Data were quoted as mean  SEM of quadruplicate determinations.

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

Starch/Stärke 2013, 65, 509–516 515

produced the lowest concentration of butyrate during fer- SCFA, particularly butyric acid, followed by the fermenta-
mentation (Table 3). This result is consistent with the report tion of RS from uncooked HAS. Their fermentation was
from Saito et al. [25] studied in an animal trial. For samples characterized by the highest production of total SCFA, and
in group 4, starches were hydrothermally treated for differ- the greatest ratios of butyrate/acetate and butyrate/total
ent times (once, twice and three times, respectively), and SCFA. Structural analyses by FTIR suggested that
their fermentation data (Table 3) showed that butyrate these two sources of RS have a higher absorption ratio
production was greatly enhanced with increasing treat- at 1047/1022 cm1 indicating a higher organized structure
ment times. because of removing less-organized region in starch by
Fermentation of RS from NTT sample produced the digestion. The analysis of molecular structure by HPLC
greatest propionate concentration as well among all showed that these two sources of RS contain larger
the samples (Table 3). Compared to acetate, propionate molecular composition (amylopectin fraction) in their
concentration was much lower, which might be due to molecular profiles, which was assumed to be one of key
only a limited number of gut bacterial species producing factors promoting butyric/SCFA production.
propionate. Many of the known propionate producing bac-
teria included Selenomonas, Mitsuokella, Megamonas, The authors wish to thank Nutrition Analysis Lab staff at
Megasphaera, Veillonella etc., which are capable of utiliz- CSIRO Human Nutrition, Adelaide, Australia, particularly
ing lactate and converting it largely to acetate and propi- Michelle Vuaran for her excellent work conducting fer-
onate. Nevertheless, in some studies it was also found that mentation experiment and SCFA analysis. Special thanks
the fermentation of RS could promote the formation of to Drs Anthony Bird & David Topping for the discussion of
propionic acid [26, 27]. the project plan, data interpretation and reviewing the
Most importantly, ratios of butyrate/acetate and buty- manuscript.
rate/total SCFA were greatly enhanced during the fermen-
tation of RS from NTTsample, followed by the fermentation The authors have declared no conflict of interest.
of RS from uncooked HAS. The much lower in those two
ratios (Table 3) obtained from the fermentation of RS from
HTT groups suggested that starch which underwent differ-
5 References
ent gelatinization/retrogradation process could modulate
bacterial fermentation pattern, subsequently influencing [1] Asp, N.-G., Resistant starch – an update on its physiological
metabolites of ferments (e.g. SCFA). effects. Adv. Exp. Med. Bio. 1997, 427, 201–210.
Furthermore, the analysis of molecular profile by HPLC [2] Bird, A. R., Brown, I. L., Topping, D. L., Starches, resistant
showed that the most pronounced difference in molecular starches, the gut microbiota and human health. Curr. Issues
Intest. Microbiol. 2000, 1, 25–37.
structure of RS between the groups favoring greater buty-
rate production (i.e. NTT sample and uncooked HAS [3] Ferguson, L. R., Tasman-Jones, C., Englyst, H., Harris, P. J.,
Comparative effects of three resistant starch preparations on
sample) and the groups supporting a lower butyrate pro- transit time and short-chain fatty acid production in rats. Nutr.
duction (i.e. HTT samples, both group 3 and group 4) was Cancer. 2000, 36, 230–237.
that the amylopectin fraction still existed in the molecular [4] Henningsson, A. M., Margareta, E., Nyman, G. L., Björck,
profile of former groups (although it is not a dominant I. M., Influences of dietary adaptation and source of
molecular fraction in the profile) but absent in the molecular resistant starch in the hindgut of rats. Br. J. Nutr. 2003,
89, 319–328.
profile of the latter groups. It seems that molecular struc-
[5] Bird, A. R., Vuaran, M., Crittenden, R., Hayakawa, T. et al.,
ture of RS plays an important role manipulating gut bac- Comparative effects of a high-amylose starch and a fructoo-
terial fermentation behavior, subsequently influencing ligosaccharide on fecal bifidobacteria numbers and short-
SCFA production in amount and proportion. chain fatty acids in pigs fed bifidobacterium animalis. Digest.
Dis. Sci. 2009, 54, 947–954.
[6] Burkitt, D. P., Some diseases characteristic of modern
Western civilization. Br. Med. J. 1973, 1, 274–278.
4 Conclusions
[7] Topping, D. L., Clifton, P. M., Short-chain fatty acids and
human colonic function: roles of resistant starch and non-
Fermentation of RS by human gut bacteria resulted in starch polysaccharides. Physiol. Rev. 2001, 81, 1031–1064.
the production of SCFA and a concomitant decline [8] Heerdt, B. G., Houston, M. A., Augenlicht, L. H., Short-chain
in pH. A significant correlation was observed between fatty acid-initiated cell cycle arrest and apoptosis of colonic
the pH value of batch culture and the net production epithelial cells is linked to mitochondrial function. Cell Growth
Differ. 1997, 8, 523–532.
of total SCFA after the fermentation. This study indicated
gut microbiota fermentation patterns can be modulated [9] Medina, V., Edmonds, B., Young, G. P., James, R. et al.,
Induction of caspase-3 protease activity and apoptosis by
by RS structure. RS from NTT sample seems to be the butyrate and trichostatin A (inhibitors of histone deacetylase):
most favorable fermentation substrate for the production of dependence on protein synthesis and synergy with a

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]

516 Z. Zhou et al. Starch/Stärke 2013, 65, 509–516

mitochondrial/cytochrome c-dependent pathway. Cancer by human faecal bacteria. J. Sci. Food Agric. 1996, 71, 209–
Res. 1997, 57, 3697–3707. 217.
[10] Nugent, A. P., Health properties of resistant starch. Nutr. Bull. [19] Cummings, J. H., Macfarlane, G. T., The role of internal
2005, 30, 27–54. bacteria in nutrient metabolism. Clin. Nutr. 1997, 16, 3–11.
[11] Rafter, J., Bennett, M., Caderni, G., Clune, Y. et al., Dietary [20] VanSoest, J. J. G., Tournois, H., deWit, D., Vliegenthart,
synbiotics reduce cancer risk factors in polypectomized and J. F. G., Short-range structure in (partially) crystalline potato
colon cancer patients. Am. J. Clin. Nutr. 2007, 85, 488–496. starch determined with attenuated total reflectance Fourier-
transform IR spectroscopy. Carbohydr. Res. 1995, 279, 201–
[12] Kihara, M., Sakata, T., Fermentation of dietary carbohydrates
214.
to short-chain fatty acids by gut microbes and its influence
on intestinal morphology of a detritivorous teleost tilapia [21] Sevenous, O., Hill, S. E., Farhat, I. A., Mitchell, J. R.,
(Oreochromis niloticus). Comp. Biochem. Phys. 1997, Organisation of the external region of the starch granules
118A, 1201–1207. as determined by infrared spectroscopy. Int. J. Biol.
Macromol. 2002, 31, 79–85.
[13] Hylla, S., Gostner, A., Dusel, G., Anger, H. et al., Effects of
resistant starch on the colon in healthy volunteers: possible [22] Liu, Q., Charlet, G., Yelle, S., Arul, J., Phase transition in
implications for cancer prevention. Am. J. Clin. Nutr. 1998, potato starch-water system. I. Starch gelatinization at high
67, 136–142. moisture level. Food Res. Int. 2002, 35, 397–407.
[14] Ferguson, M. J., Jones, G. P., Production of short-chain [23] Zhou, Z. K., Topping, D. L., Morell, M., Bird, A. R., Changes in
fatty acids following in vitro fermentation of saccharides, starch physical characteristics following digestion of foods in
saccharide esters, fructo-oligosaccharides, starches, modi- the human small intestine. Br. J. Nutr. 2010, 104, 573–581.
fied starches and non-starch polysaccharides. J. Sci. Food
[24] Bird, A. R., Topping, D. L., in: Cho, S. S. Dreher, M. L. (Eds.),
Agric. 2000, 80, 166–170.
Handbook of Dietary Fibre, Marcel Dekke, New York 2001,
[15] Zampa, A., Silvi, S., Fabiani, R., Morozzi, G. et al., Effects of pp. 147–158.
different digestible carbohydrates on bile acid metabolism
[25] Saito, K., Ito, T., Kuribayashi, T., Mochida, K. et al., Effect of
and SCFA production by human gut micro-flora grown in an in
raw and heat-moisture treated high-amylose corn starch on
vitro semi-continuous culture. Anaerobe 2004, 10, 19–26.
fermentation by the rat cecal bacteria. Starch/Stärke 2001,
[16] Edwards, C. A., Rowland, I. R., in: Schweizer, T. F. Edwards, 53, 424–430.
C. A. (Eds.), Dietary Fibre – A Component of Food, Springer
[26] Andrieux, C., Pacheco, E. D., Bouchet, B., Gallant, D., Szylit,
Verlag, London 1992, pp. 119–136.
O., Contribution of the digestive tract microbiota to amy-
[17] Nordgaard, I., Mortensen, P. B., Langkilde, A. M., Small lomaize starch degradation in the rat. Br. J. Nutr. 1992,
intestinal malabsorption and colonic fermentation of resistant 67, 489–499.
starch and resistant peptides to short-chain fatty acids.
[27] Lopez, H. W., Levrat-Verny, M. A., Coudray, C., Besson, C.
Nutrition 1995, 11, 129–137.
et al., Class 2 resistant starches lower plasma and liver lipids
[18] Edwards, C. A., Gibson, G., Champ, M., Jensen, B. B. et al., and improve mineral retention in rats. J. Nutr. 2001, 131,
In vitro method for quantification of the fermentation of starch 1283–1289.

ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim [Link]