npj | science o ood A icle
Published in pa ne ship wi h Beijing Technology and Business Uni e si y
h ps://doi.o g/10.1038/s41538-024-00320-8
Tailo ing he na u al a e suga s
D- aga ose and L-so bose o p oduce
no el unc ional ca bohyd a es
Check o upda es
Oswaldo He nandez-He nandez 1, Ca los Saba e 2,3, Inés Cal e e-To e2,3,ElisaG.Doyagüez
4,
Ana M. Muñoz-Lab ado 1,C is inaJulio-Gonzalez
1, Blanca de las Ri as5, Rosa io Muñoz5,
Lo ena Ruiz 2,3, Abela do Ma golles2,3, José M. Mancheño6&F.Ja ie Mo eno 1
This mul idisciplina y s udy de ails he biosyn hesis o no el non-diges ible oligosaccha ides de i ed
om a e suga s, achie ed h ough ans uc osyla ion o D- aga ose and L-so bose by le ansuc ase
om Bacillus sub ilis CECT 39 (SacB). The cha ac e iza ion o hese ca bohyd a es using NMR and
molecula docking was ins umen al in elucida ing he ca aly ic mechanism and subs a e p e e ence
o SacB. Taga ose-based oligosaccha ides we e highe in abundance han L-so bose-based
oligosaccha ides, wi h he mos ep esen a i e s uc u es being: β-D-F u-(2→6)-β-D-F u-(2→1)-D-
Tag and β-D-F u-(2→1)-D-Tag. In i o s udies demons a ed he esis ance o aga ose-based
oligosaccha ides o in es inal diges ion and hei p ebio ic p ope ies, p o iding insigh s in o hei
s uc u e- unc ion ela ionship. β-D-F u-(2→1)-D-Tag was he mos esis an s uc u e o small-
in es inal diges ion a e h ee hou s (99.8% emained unal e ed). This disaccha ide and he
comme cial FOS clus e ed in simila b anches, indica ing compa able modula o y p ope ies on
human ecal mic obio a, and exe ed a highe bifidogenic e ec han unmodified aga ose. The
biocon e sion o selec ed a e suga s in o β- uc osyla ed species wi h a highe deg ee o
polyme iza ion eme ges as an e ficien s a egy o enhance he bioa ailabili y o hese ca bohyd a es
and p omo e hei in e ac ion wi h he gu mic obio a. These findings open up new oppo uni ies o
ailo ing na u al a e suga s, like D- aga ose and L-so bose, o p oduce no el biosyn hesized
ca bohyd a es wi h unc ional and s uc u al p ope ies desi able o use as eme ging p ebio ics and
low-calo ie swee ene s.
Acco ding o he In e na ional Socie y o Ra e Suga s (ISRS, h ps://www.
is s.kagawa-u.ac.jp/), a e suga s a e monosaccha ides and hei de i a i es
ha a esca ceinna u e
1. Main g oups o a e suga s include ke ohexoses
such as aga ose, so bose, and psicose (in bo h D- and L-configu a ions),
polyols like xyli ol, and deoxygena ed monosaccha ides such as L- ibose.
Despi e hei low na u al abundance, a e suga s ha e significan po en ial
o p ac ical applica ions in he ood, pha maceu ical, and nu i ion
indus ies due o hei biological unc ions2–4. The Izumo ing s a egy,
de eloped by P o . Izumo i a he Ra e Suga Resea ch and Educa ion
Cen e o Kagawa Uni e si y in Japan, has been a leading app oach o
syn hesizing a esuga s o wen yyea s.Thisme hod elieson he p inciple
ha monosaccha ides can be cyclically con e ed h ough enzyma ic p o-
cesses, including epime iza ion, isome iza ion, and oxida ion- educ ion5,6.
Howe e , ecen ly, se e al non-Izumo ing enzyma ic echniques ha e
eme ged, based on di e en p inciples such as aldose epime iza ion, enzy-
ma ic condensa ion, phospho yla ion-dephospho yla ion cascade eac-
ions, dica boxylic eac ions, and he enzyma ic syn hesis o no el
disaccha ides using a e suga s as accep o s7.No ably,onlya ewpape s
ha e add essed he la e app oach, mainly ocusing on he xylosyla ion o
D-psicose (also known as D-allulose)6, o he glucosyla ion o D-galac ose,
1Ins i u e o Food Science Resea ch, CIAL (CSIC-UAM), Nicolas Cab e a 9, 28049 Mad id, Spain. 2Dai y Resea ch Ins i u e o As u ias (IPLA-CSIC), Paseo Río
Lina es s/n, 3300 Villa iciosa, As u ias, Spain. 3Heal h Resea ch Ins i u e o As u ias (ISPA), A enida Hospi al Uni e si a io s/n, 33011 O iedo, As u ias, Spain.
4Cen o de Química O gánica “Lo a Tamayo”(CSIC), Juan de la Cie a 3, 28006 Mad id, Spain. 5Ins i u e o Food Science, Technology and Nu i ion, ICTAN (CSIC),
Juan de la Cie a 3, 28006 Mad id, Spain. 6Ins i u e o Physical Chemis y ‘Blas Cab e a’(IQF-CSIC), Se ano 119, 28006 Mad id, Spain.
e-mail: [email p o ec ed]
npj Science o Food | (2024) 8:74 1
1234567890():,;
1234567890():,;
D- ibose, L-a abinose, o D-psicose ca alyzed by bac e ial suc ose
phospho ylases8,9. Fuji a e al. 10 also desc ibed he ans uc osyla ion o
L-so bose o o masingledisaccha idewi h heglycosidiclinkageβ-(2 →2)
ca alyzed by a β- uc o u anosidase om Mic obac e ium saccha ophilum
K-1 ( o me ly known as A h obac e sp. K-1). These s udies sugges ha
ca bohyd a es con aining a e suga s could ha e in e es ing physiological
ac i i ies and po en ially highe unc ionali y han he a e suga s hem-
sel es. Howe e , as Zhang e al. 11 highligh ed, he e is a lack o li e a u e on
he specific physiological unc ions o oligosaccha ides con aining a e
suga s, making his a key a ea o u u e esea ch in a e suga s udies.
D-Taga ose sha es physical p ope ies wi h suc ose in e ms o colo ,
ex u e, and swee ness, bu appea s o ha e a much lowe glycemic index
and calo ic alue due o i s low bioa ailabili y12.Asa esul ,i s p ope iesas a
low-calo ie swee ene make D- aga ose he mos demanded and comme -
cially impo an a e suga . The U.S. Food and D ug Adminis a ion (FDA)
ca ego ized D- aga ose as a Gene ally Recognized as Sa e (GRAS) subs ance
wo decades ago11. Subsequen ly, i was au ho ized as a no el ood in
Aus alia and New Zealand13, and in he Eu opean Union, based on an
opinion om he UK Ad iso y Commi ee o No el Foods and P ocesses
(ACNFP)14. The use o D- aga ose in ood p oduc s and be e ages is also
accep ed in o he coun ies, including Sou h A ica and Sou h Ko ea15,16.
Mo eo e , aga ose has shown po en ial beneficial e ec s and he apeu ic
p ope ies in humans and has been p oposed o ea ing condi ions like
“ ype 2”diabe es, hype glycemia, anemia, and hemophilia, as well as o
p omo ing gu heal h h ough i s p ebio ic p ope ies11,17. Howe e , he ole
o aga ose as a p ebio ic ing edien could be unde mined by s udies
showing ha D- aga ose is abso bed a leas o some ex en in he small
in es ine in a s18 and o a g ea e ex en in humans19. A s udy in humans
indica ed a mean abso p ion a e o 80% ( ange 69–88%) in he small
in es ine and a u ina y exc e ion o ei he 1% o 5%, so only he emaining
20% o D- aga ose could be e men ed in he colon. These alues we e
accep ed and used by he Eu opean Food Sa e y Au ho i y (EFSA) o es i-
ma e heene gygain omabso bedandme abolizedD- aga ose20.Basedon
hese findings, he use o enzymes wi h ansglycosidase ac i i y o ailo
aga ose h ough selec i e elonga ion could yield no el ca bohyd a es wi h
highe unc ionali y han aga ose i sel . Specifically, he enzyma ic p o-
duc ion o a disaccha ide esul ing om aga ose uc osyla ion could no
only mimic he physical p ope ies o suc ose bu also o e significan
efinemen . This is because such a disaccha ide would likely esis gas o-
in es inal diges ion and pass h ough he small in es ine, becoming a ailable
o consump ion by he colonic mic obio a o a g ea e ex en han aga ose,
and wi h di e en selec i i y due o key s uc u al modifica ions.
The e o e, his a icle epo s he biosyn hesis o no el non-
diges ible aga ose-based oligosaccha ides h ough a ans uc osyla-
ion eac ion ca alyzed by a le ansuc ase om Bacillus sub ilis CECT
39 (SacB). Addi ionally, o he a e suga s such as L-so bose and
D-psicose we e es ed as po en ial accep o s o p oducing no el
ca bohyd a es, while molecula docking u he e ified he easibili y
o hese subs a es and enzyme binding. A comp ehensi e s uc u al
cha ac e iza ion o he biosyn hesized D- aga ose- and L-so bose-
oligosaccha ides was accomplished by NMR. In i o s udies
demons a ing he esis ance o in es inal diges ion and p ebio ic
p ope ies o he main aga ose-based disaccha ide p o ided insigh s
in o i s s uc u e- unc ion ela ionship.
Resul s and Discussion
Accep o specifici y o SacB
Le ansuc ases (also known as suc ose:2,6-β-D- uc an 6-β-D- uc-
osyl ans e ases o be a-2,6- uc osyl ans e ases, EC 2.4.1.10) ca a-
lyze bo h suc ose hyd olysis and he β-(2,6)-linked le an
polyme iza ion. Likewise, in he p esence o app op ia e subs a e
accep o s o he han wa e le ansuc ases a e among he mos used
and e ec i e mic obial uc ansuc ases in oligosaccha ide syn hesis
om suc ose21,22.The‘ca bohyd a e-ac i e enzymes’da abase (CAZY,
h p://www.cazy.o g/)23 g oups bac e ial le ansuc ases in o glycoside
hyd olase (GH) amily 68. The le ansuc ase SacB om Bacillus
sub ilis (h ps://www.unip o .o g/unip o /P05655;h ps://www.ncbi.
nlm.nih.go /p o ein/P05655.1) is likely one o he mos s udied
uc ansuc ases. I consis s o a 53 kDa (including he signal pep ide)
single-domain p o ein wi h a fi e-bladed p opelle old ha encloses
a subs a e-binding cen al ca i y (i.e., ac i e si e) loca ed in a deep
pocke su ounded by se e al semi-conse ed amino acid esidues24.
The ca aly ic mechanism o SacB is based on he double-
displacemen eac ion mechanism wi h e en ion o he anome ic
configu a ion o he ans e ed uc osyl moie ies coo dina ed by
h ee ca aly ic esidues (Asp86, Asp247, and Glu342)24,25.B iefly, he
ca aly ic mechanism in ol es he p o ona ion o he glycosidic oxy-
gen by a ca boxyla e ac ing as a gene al acid ca alys , mos p obably
Glu34224, concu en ly wi h clea age o he glycosidic bond in a
uc osyl dono (usually a suc ose molecule). This p ocess akes place
be ween he −1and+1 subsi es, acco ding o he nomencla u e
p oposed by Da ies e al. 26, wi hin he ac i e si e o SacB. This fi s
s ep esul s in he o ma ion o a uc osyl-SacB co alen in e -
media e, wi h Asp86 ac ing as he nucleophile, and he elease o he
glucose moie y. In he second s ep, he co alen ly-bound uc osyl
moie y is ans e ed om Asp86 o an accep o molecule wi h
e en ion o he anome ic configu a ion, leading o he elonga ion o
he subs a e accep o in one uc osyl uni 27. The accep o specifici y
o SacB has been explo ed o a long ime by using a ange o ca -
bohyd a es and/o de i a i es as uc osyl accep o s in eac ions
using suc ose as dono subs a e25,28–33. Howe e , o he bes o ou
knowledge, he accep o specifici y o SacB on a e suga s such as D-
aga ose, L-so bose, o D-psicose has ne e been ackled.
Figu e 1shows he GC-FID p ofiles esul ing om he enzyma ic
eac ion mix u es ca alyzed by SacB using as s a ing subs a es D-
suc ose:D- aga ose and D-suc ose:L-so bose a a concen a ion a io o 200:
200 g L−1(mole a io o D-suc ose o L-so bose o D- aga ose is1 :1.9). Bo h
D- aga ose and L-so bose seemed o be good accep o s since new p oduc s
ha ing elu ion imes compa ible wi h he p esence o ca bohyd a es wi h a
deg ee o polyme iza ion om 2 o 5 could be de ec ed (Fig. 1A, B), whe eas
he p esence o D-psicose did no lead o he o ma ion o de ec able new
p oduc s (da a no shown). D-Taga ose and L-so bose a y in wo posi ions
Fig. 1 | GC-FID p ofiles o he eac ion p oduc s ca alyzed by he le ansuc ase
SacB om Bacillus sub ilis CECT 39. S a ing subs a es: (A) D-Suc ose:D-Taga-
ose, (B) D-Suc ose:L-So bose bo h a a concen a ion a io o 200: 200 g L−1. DP:
deg ee o polyme iza ion. DP2: Disaccha ide ac ion; DP3: T isaccha ide ac ion;
DP4: Te asaccha ide ac ion; DP5: Pen asaccha ide ac ion.
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 2
a C-4 andC-5 si es o he placemen o he-OH g oup, whe easD- aga ose
and D-psicose ha e hese a ia ions a C-3 and C-4 si es, and L-so bose and
D-psicose a C-3 and C-5 si es (Fig. S1). Acco ding o he ca aly ic
mechanism o SacB explained abo e, we would expec he o ma ion o
disaccha ides esul ing om he ans e o a uc osyl esidue om suc ose
o any o he hyd oxyl g oups a ailable o D- aga ose and L-so bose. Like-
wise, he new p oduc s ha ing a deg ee o polyme iza ion abo e 2 could be
de i ed om u he ans uc osyla ion eac ions, esul ing in chain elon-
ga ion and oligome o ma ion.
In he nex subsec ion, he comp ehensi e s uc u al cha ac e iza ion
accomplished by NMR o he biosyn hesized D- aga ose- and L-so bose-
ca bohyd a es is desc ibed. Fu he mo e, in silico molecula docking
expe imen s o u he in es iga e he p ecise mechanism o o de ed-
subs a e binding o SacB we e ca ied ou o e i y he easibili y o hese
s uc u al complexes o SacB wi h he es ed a e suga s and o gain
knowledge on i s subs a e specifici y.
S uc u al elucida ion o D- aga ose and L-so bose-
uc osyla ed ca bohyd a es
NMRcha ac e iza ionwasaccomplishedby hecombineduseo 1Dand2D
[1H, 1H] and [1H, 13C] NMR expe imen s (gCOSY, TOCSY, mul iplici y-
edi ed gHSQC, gHMBC and hyb id expe imen gHSQC-TOCSY). 1Hand
13C NMR chemical shi s obse ed a e summa ized in Tables S1–S6. A ull
se o spec a is also a ailable in he Suppo ing In o ma ion (Fig-
u es S2–S29).
D- aga ose De i a i es: D- aga ose de i a i es was di ided among
h ee pu ified ac ions, labeled T1, T2, and T3.
Pu ified F ac ion T1: Analysis o pu ified ac ion T1 e ealed
he p esence o wo anome s o a disaccha ide s uc u e, iden ified as
β-D- uc o u anosyl-(2 →1)-D- aga opy anose, exis ing in a 6:1 α:β
a io (Fig. 2A,B).No ably, he1D1H NMR spec um o his
compound displayed an absence o signals in he anome ic egion,
a ibu ed o he lack o anome ic p o ons wi hin he uc ose and
aga ose uni s. Fu he mo e, he 1D 13C NMR spec um e ealed
signals o 24 ca bons, encompassing ou anome ic ca bons, whose
chemical shi s we e ins umen al in es ablishing he py anose and
u anose o ms o aga ose and uc ose, espec i ely34. A de ailed
analysis using mul iplici y-edi ed gHSQC spec um acili a ed he
linkage o ca bon signals o hei co esponding p o on esonances,
he eby ein o cing he s uc u al iden ifica ion. The glycosidic lin-
kages we e u he analyzed and confi med h ough gHMBC spec-
um obse a ions, he eby a fi ming he (2 →1) linkage be ween he
cons i uen suga s. The assignmen o α-andβ-anome s o each
compound was suppo ed by 1Hand13C NMR shi s (Table S1)34.
Pu ified F ac ion T2: F ac ion T2 p esen ed a di e gen disaccha ide
s uc u e, de oid o aga ose, iden ified as β-D- uc o u anosyl-(2 →6)-D-
glucopy anose. This s uc u e was elucida ed h ough a simila comp e-
hensi e sui e o NMR analyses, including 1D and 2D spec a, which sup-
po ed he p esence o β-andα-glucose alongside wo β- uc ose uni s. The
assignmen o glycosidic linkages was co obo a ed h ough de ailed
gHMBC spec al analysis, confi ming he (2 →6) linkage. These esul s a e
in acco d wi h o he da a om Sa o e al. 35.
Pu ified F ac ion T3: In ac ion T3, h ee isaccha ides we e iden i-
fied, wi h he majo compound being a complex β-D- uc o u anosyl-
(2 →6)-β-D- uc o u anosyl-(2 →1)-D- aga opy anoses uc u e(Fig.2C,
D), alongside ano he isaccha ide in lesse p opo ion and a e y mino
signal indica i e o a po en ial hi d s uc u e, assigned as β-D- uc o u -
anosyl-(2 →6)-α-D-glucopy anosyl-(1 →2)-β-D- uc o u anoside. The
NMR da a, including 1D 13C NMR and mul iplici y-edi ed gHSQC spec a,
p o ided conclusi e e idence o he p oposed s uc u es, wi h glycosidic
linkages p ecisely de e mined h ough gHMBC spec al analysis, and in
acco d wi h o he da a ega ding (2 →6) linked uc oses om Jakob e al.,
36 and Tajima e al. 37.
L-so bose De i a i es. Fou dis inc ac ions we e analyzed, e ealing
di e en compounds wi hin each ac ion.
Pu ified F ac ion S1: F ac ion S1 o e ed a compelling mix u e, wi h
suc ose iden ified as he majo componen alongside α-and
β-glucopy anose. No ably, his ac ion ea u ed significan compounds
con aining L-so bose: he disaccha ide β-D- uc o u anosyl-(2 →5)-α-L-
so bopy anose and he isaccha ide β-D- uc o u anosyl-(2 →6)-β-D-
uc o u anosyl-(2 →5)-α-L-so bopy anose, as depic ed in Figu e S14
(compounds 6 and 7, espec i ely). The disaccha ide, he eina e e e ed o
as β-D-F u-(2 →5)-α-L-So , showcased h ough he 1D 13C NMR spec-
um, exhibi ed signals o 12 ca bons, highligh ing he p esence o only he
Fig. 2 | S uc u es accomplished by NMR o di-
and isaccha ides ob ained by aga ose uc osy-
la ion ca alyzed by he le ansuc ase SacB om
Bacillus sub ilis CECT 39. A β-D- uc o u anosyl-
(2→1)-α-D- aga opy anose. Bβ-D- uc o u -
anosyl-(2→1)-β-D- aga opy anose. Cβ-D- uc o-
u anosyl-(2 →6)-β-D- uc o u anosyl-(2 →1)-α-
D- aga opy anose. Dβ-D- uc o u anosyl-(2 →6)-
β-D- uc o u anosyl-(2 →1)-β-D- aga opy anose.
O
H
HO H
HOH
O
HO
OH
O
H
HO
H
HO
OH
H
H
OH O
H
HO H
HOH
O
HO
OH
O
H
HO
H
HO
OH
H
H
OH
O
H
HO H
HOH
O
OH
O
H
HO
H
HO
OH
H
H
OH
O
H
HO
H
H
OH
O
HO
OH
O
H
HO H
HOH
O
OH
O
H
HO
H
HO
OH
H
H
OH
O
H
HO
H
H
OH
O
HO
OH
AB
CD
F uII F uIF uII F uI
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 3
α-anome o so bose. This included wo anome ic ca bons indica ing he
py anose o m o α-so bose and he β- u anose o m o uc ose38,39,a
s uc u e suppo ed by mul iplici y-edi ed gHSQC spec um ha linked
ca bon signals o co esponding p o on esonances. Va ious NMR expe i-
men s, including COSY, TOCSY, and gHSQC-TOCSY, co obo a ed hese
findings, wi h gHMBC spec um analysis u he confi ming he (2 →5)
linkage. 1Hand13C chemical shi s o C-1 and H-4 o β-F u , and C-5 and
H-5 o so bose also suppo ed he p oposed s uc u es by compa ison wi h
o he da a desc ibed in li e a u e35,36,40,41.
F ac ion S2: Ad ancing o ac ion S2, he in es iga ion iden ified he
same β-D-F u-(2 →5)-α-L-So as a majo componen , alongside new i-
saccha ides. No ably, one isaccha ide mi o ed he s uc u e o he majo
compound om ac ion S1, β-D-F u-(2 →6)-β-D-F u-(2 →5)-α-L-So ,
ou lined in Fig. S18 (compound 7). Addi ional compounds in his ac ion
included β-D- uc o u anosyl-(2 →6)-α-D-glucopy anosyl-(1 →2)-β-D-
uc o u anoside ( he same as ound in ac ion T3, depic ed in Fig. S10,
compound 4) and β-D- uc o u anosyl-(2 →3)-α-D-glucopy anosyl-
(1 →2)-β-D- uc o u anoside (Figu e S18, compound 8).
F ac ions S3 and S4: In ac ion S3, he na a i e ex ends wi h he
iden ifica ion o wo e asaccha ides. One, deno ed as compound 9 (Fig.
S22),se es asanex ensiono he isaccha ides uc u e om ac ionS1by
including an addi ional β- uc o u anose uni , illus a ing he con inued
(2 →6) linkage. The o he , compound 10 (Fig. S22), ollows a simila
ex ension pa e n bu eplaces L-so bose wi h glucose. F ac ion S4 un eils a
pen asaccha ide. This s uc u e, depic ed in Fig. S26 (compound 11),
ep esen s a u he elonga ion o e asaccha ide 9 wi h an addi ional
β- uc o u anose uni , main aining he consis en (2 →6) linkage pa e n.
Molecula docking s udies
Molecula docking we e ca ied ou using he s uc u e o SacB (PDB en y:
1PT2) and he s uc u es o he a e suga s D- aga ose, L-so bose, and
D-psicose. These suga s uc u es we e e ie ed om he PubChem da a-
base and e ified using eLBOW om he Phenix package42, which allows o
he geome y op imiza ion o ligands. As indica ed abo e, hese h ee suga s
di e in heabsolu econfigu a ion o wo ca bon a oms in pai wise com-
pa isons (Fig. S1). To p oceed wi h he molecula docking wi h Au oDock
Vina and since ou in e es esides in he simula ion o he uc osyl ans e
s ep o he a e suga s D- aga opy anose, L-so bopy anose and
D-psicopy anose as accep o s, he ecep o o he molecula docking wi h
Au oDock Vina was a co alen uc osyl-enzyme in e media e whe e he
Asp86 side chain o SacB is bound o he C2 a om o he uc osyl uni . The
configu a ion o his la e C2 a om is heopposi eo heonep esen in he
suc ose molecule in ag eemen wi h he e en ion o he configu a ion
ca aly ic mechanism o SacB24, in which he Asp86 side chain ac s as a
nucleophile25,43. Since no u he changes we e made ega ding cha ges o
he uc osyl a oms o he pu a i e co alen uc osyl-enzyme in e media e,
hese in silico esul s should be mainly in e p e ed in quali a i e e ms.
None heless, we belie e ha he ob ained poses o he a e suga s wi hin he
ac i e si e o SacB hus p epa ed p o ide a solid guide o in e p e ing in
molecula e ms he syn hesis o he disaccha ides β-D- uc o u anosyl-
(2→1)-D- aga opy anose and β-D- uc o u anosyl-(2-5)-α-L-so bopy -
anose and he absence o uc osyl ans e o D-psicopy anose by SacB.
Figu e 3illus a es he binding ea u es o he h ee a e suga s D-
aga opy anose, L-so bopy anose, and D-psicopy anose wi h he ecep o
SacB. Fi s , h ee main conside a ions can be made be o e desc ibing he
de ailed suga -p o ein in e ac ions. Fi s ly, he e is only one pose o
D- aga opy anose and o L-so bopy anose compa ible wi h he o ma ion
o he disaccha ides β-D- uc o u anosyl-(2-1)-D- aga opy anose and β-D-
uc o u anosyl(2-5)-α-L-so bopy anose epo ed in his wo k; secondly,
hese wo poses ag ee wi h he ca aly ic mechanism p oposed o he uc-
osyl ans e eac ion by SacB25,43; hi dly, he eisonepose o
D-psicopy anose which is almos supe imposable wi h L-so bopy anose.
Rema kably, he s e eochemis y o he C5 a om o D-psicopy anose, which
di e s om ha o L-so bopy anose, is no compa ible wi h he p og ess o
he ca alysis o he uc osyl ans e .
Rega ding he binding mode o D- aga opy anose (Fig. 3), possible H
bonds a e o med be ween he hyd oxyl g oup -OH2 and he OE1 and OE2
a oms o Glu340, be ween he -OH3 g oup and he NH2 a om o A g246
and OE2 a om o Glu342, and be ween he -OH4 g oup and he NH2 a om
o A g246. The -OH1 hyd oxyl g oup is a H bond dis ance o he NE a om
o A g360 and impo an ly o he OE1 and OE2 a oms o he gene al acid-
base ca aly ic esidue Glu342. Mos p obably, hese la e in e ac ions
p ope ly o ien a e his hyd oxyl g oup owa ds he C2 a om o he uc osyl
uni co alen ly bound o Asp86 o he p oduc i e nucleophilic a ack,
which would explain he o ma ion o β-D- uc o u anosyl(2-1)-β-D-
aga opy anose.
Con e sely, binding o L-so bopy anose (Fig. 3)wouldin ol e he
in e ac ion be ween he -OH1 wi h he NH2 a om o A g246, wi h he ND2
and OD1 a oms om Asn242, and wi h he OE2 a om om Glu340,
be ween he -OH3 hyd oxyl g oup wi h he NH2 a om o A g246 and
finally, be ween he -OH4 hyd oxyl g oup wi h he OE2 om Glu342 and
wi h he NH2 a om o A g246. Simila o he -OH1 g oup o D-
Fig. 3 | Molecula docking o he binding modes o he a e suga s es ed in his
s udy (i.e., D- aga opy anose, L-so bopy anose, and D-psicopy anose) o he
pu a i e co alen uc osyl in e media e o le ansuc ase om Bacillus sub ilis
CECT 39 (SacB). . Bo h he suga s and ep esen a i e amino acid side chains om
he SacB ac i e si e a e shown as s icks model; po en ial H bonds a e indica ed by
black dashed lines and pu a i e in e ac ions o -OH1 om D- aga opy anose o
-OH5 om L-so bopy anose a e shown as cyan dashed lines. The dis inc s e eo-
chemis y o C5 o D-psicopy anose does no pe mi he uc osyla ion eac ion o
p oceed (see he ex o u he de ails).
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 4
aga opy anose, he -OH5 g oup o L-so bopy anose in e ac s wi h he NE
a om o A g360 and wi h he OE1 and OE2 a oms o Glu342, p ope ly
o ien ing his g oup owa ds he C2 a om o he uc osyl uni . When he
poses o hese wo suga s a e compa ed, i can be obse ed ha he e a e
pai s o equi alen -OH g oups ega ding hei in e ac ions wi h SacB: he
-OH1, -OH3, and -OH4 hyd oxyl g oups om D- aga opy anose would be
equi alen o he -OH5, -OH4 and -OH3 om L-so bopy anose, espec-
i ely. Mos p obably he hyd oxyl g oups -OH5 om D- aga opy anose
and -OH2 om L-so bopy anose would in e ac wi h p o ein ligands
h ough sol en molecules.
Finally, as indica ed abo e, he binding mode o D-psicopy anose is
e y simila o ha o L-so bopy anose wi h he hyd oxyl g oups -OH1,
-OH2, and -OH3 occupying equi alen posi ions as he -OH1, -OH2, and
-OH4 g oups om L-so bopy anose (Fig. 3). Rema kably, he o ien a ion
(and p ecise posi ion) o he -OH5 g oup o D-psicopy anose clea ly di e s
om he -OH5 om L-so bopy anose due o hei di e en s e eo-
chemis y: axial o D-psicopy anose and equa o ial o L-so bopy anose.
Now, he -OH5 g oup om D-psicopy anose poin s owa ds he acidic
g oups Glu340 and Glu342, whe eas he -OH5 g oup om
L-so bopy anose (o -OH1 om D- aga opy anose), in close p oximi y o
Glu342, is di ec ed owa ds he uc osyl uni . Al oge he hese in silico
esul s p o ide a s uc u al basis o he o ma ion o he disaccha ides β-D-
uc o u anosyl-(2→1)-D- aga opy anose and β-D- uc o u anosyl-(2-5)-
α-L-so bopy anose and also p o ide a plausible explana ion o he lack o
uc osyl ans e o D-psicopy anose.
Enzyma ic syn hesis op imiza ion o D- aga ose- uc osyla ed
ca bohyd a es
Based on he widesp ead use and significance o D- aga ose as a low-calo ie
swee ene wi h po en ial p ebio ic p ope ies, an op imiza ion was con-
duc ed ocusing on he enzyma ic eac ions using D- aga ose as he s a ing
subs a e. This op imiza ion ook in o accoun he SacB enzyme con-
cen a ion and he dono :accep o concen a ion a ios.
Ini ially, he concen a ion o SacB was op imized in eac ion mix u es
wi h ini ial concen a ions se a 200:200 g L−1o D-suc ose:D- aga ose.
Among he h ee es ed SacB concen a ions (0.6, 3.1, and 6.2 U mL−1), he
in e media e concen a ion o 3.1 U mL−1led o he highes con en o he
main accep o p oduc , β-D-F u-(2 →1)-D-Tag, wi h a 5- old and 13- old
inc ease in concen a ion compa ed o 6.2 and 0.6 U mL−1, espec i ely
(da a no shown).
A e se ing he enzyme concen a ion a 3.1 U/mL, he ini ial D-
suc ose:D- aga ose concen a ion a ios we e op imized using he a ios
200:200, 200:300, 200:450, and 300:300 g/L. Figu e 4displays he e olu ion
o uc ose, glucose, suc ose, and β-D-F u-(2 →1)-D-Tag in he eac ion
ca alyzed by SacB (3.1 U mL−1) o e 72 hou s a he ou di e en s a ing
concen a ions o D-suc ose:D- aga ose mix u es. In e e y ins ance, β-D-
F u-(2 →1)-D-Tag syn hesis commenced wi hin he fi s hou o eac ion,
eaching maximum o ma ion a 24 h and main aining s abili y o he
du a ion o he eac ion pe iod, signi ying he high s abili y o he main
p oduc o med. The a ious D-suc ose:D- aga ose concen a ion a ios
esul ed in di e en maximum concen a ions o β-D-F u-(2 →1)-D-Tag,
anging om 27 g L−1(wi h a 200:200 g L−1 a io) o 64 g L−1(wi h a
300:300 g L−1 a io). These esul s co obo a e ha an inc ease in he con-
cen a ion o he dissol ed subs a es leads o g ea e e ficiency in accep o
eac ions ca alyzed by glycoside hyd olases due o he educ ion o wa e
concen a ion22,33.
In e ms o con e sion pe cen ages, he 300:300 and 200:450 g L−1
concen a ion a ios es ed yielded he highes pe cen ages o 21.5% and
26.0%, espec i ely. Howe e , he a ios whe e D- aga ose was in excess
ela i e o suc ose esul ed in lowe yields o β-D-F u-(2 →1)-D-Tag,
indica ing a less e ficien sys em due o he lowe con e sion o D- aga ose
in o he p oduc . This aligns wi h p e ious s udies using his enzyme o he
Fig. 4 | E olu ion o uc ose, glucose, suc ose and β-D-F u-(2 →1)-D-Tag
(labelled as aga osyl- uc ose) ollowing he eac ion ca alyzed by he le ansu-
c ase SacB om Bacillus sub ilis CECT 39 a ou di e en s a ing suc ose- o-
aga ose concen a ion a ios (labelled as g pe 100 mL o eac ion). The
biosyn hesis o he main eac ion p oduc (i.e., β-D-F u-(2 →1)-D-Tag) eached a
maximum a 24 h and, hen, was s able o he whole eac ion pe iod (72 hou s)
ega dless o he s a ing subs a e concen a ion a io; howe e , he s a ing sub-
s a e concen a ion a io had an influence on he yield o he main eac ion p oduc .
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 5
syn hesis o he isaccha ide e lose (α-D-Glc-(1 →4)-α-D-Glc-(1 →2)-
β-D-F u) om suc ose/mal ose eac ion mix u es33.
To sum up, he op imal D-suc ose:D- aga ose concen a ion a io was
300:300 g/L, as i p oduced he highes concen a ion o β-D-F u-(2 →1)-
D-Tag (64.04 g L−1) and yield (21.3%), along wi h a con e sion pe cen age
(21.5%) compa ed o he o he es ed concen a ion a ios.
In i o diges ibili y o aga ose-based uc osyla ed ca bohy-
d a es by pig BBMV disaccha idases
Human da a indica e ha D- aga ose is abso bed in he small in es ine wi h
an a e age abso p ion a e o 80%19,20. The e o e, he biocon e sion o
D- aga ose in o uc osyla ed species wi h a highe deg ee o polyme iza ion
could be an e ec i e s a egy o inc ease he bioa ailabili y o aga ose and
enhance i s in e ac ion wi h he gu mic obio a. This would only be possible
i he uc osyla ed aga ose emains la gely una ec ed by he hyd oly ic
ac i i y o he mucosal disaccha idases in he mammalian small in es inal
b ush bo de memb ane esicles (BBMV), a key s ep in he diges ion p o-
cess o s a chy and non-s a chy ca bohyd a es44. Mo eo e , esis ance o
in es inal diges ion is a c i ical equi emen o he po en ial use o β-D-F u-
(2 →1)-D-Tag and i s de i a i es as no el, low-calo ie unc ional
ing edien s.
Asexpec ed,suc ose,whichwasusedasaposi i econ olin hein i o
diges ion sys em using pig BBMV, was eadily diges ed and was no longe
de ec able a e 1 h (Table 1). The deg ees o esis ance o β-D-F u-(2 →1)-
D-Tag and he isaccha ide β-D-F u-(2 →6)-β-D-F u-(2 →1)-D-Tag
we e de e mined ollowing he in i o diges ion o hei pu ified o ms
isola ed om he enzyma ic eac ion mix u e (included in T1 and T3
ac ions). Ou da a sugges ha hese no el aga ose-based uc osyla ed
ca bohyd a es a e highly esis an o he hyd oly ic ac ion o he mucosal
disaccha idases in pig small in es inal BBMV h oughou he en i e 3-hou
diges ionpe iod,suppo ing hei po en ialaslow-calo ieing edien s(Table
1). Inulin- ype uc ooligosaccha ides (FOS), known o hei β-(2 →1)
glycosidiclinkagesbe ween uc ose esidues,ha ep e iouslydemons a ed
simila esis ance o mammalian diges i e enzymes45,46.
The e o e, he low diges ibili y o indiges ibili y o he biosyn hesized
aga ose-basedca bohyd a escalls o u he s udieson hei e men abili y
and po en ial o selec i ely modula e he gu mic obio a.
In i o e men a ion p ope ies o β-D-F u-(2→1)-D-Tag using
human ecal samples
The modula o y e ec o he main p oduc de i ed om he ans uc o-
syla ion o aga ose, i.e.β-D-F u-(2→1)-D-Tag on he human ecal
mic obio a, was compa ed o hose o comme cial (unmodified) aga ose
andFOS.Fo hispu pose,ba ch ecalin i o e men a ionsusing ecalpool
inoculums om heal hy dono s we e ca ied ou . Aliquo s we e collec ed a
di e en e men a ion imes (0, 8, 24 h) and mic obiome composi ion
p ofiles we e de e mined by 16S RNA sequencing. Alpha di e si y coe fi-
cien s we e fi s calcula ed o assess he a iabili y o gene a wi hin ecal
homogena es ( ecal pool inoculums co esponding o he ini ial
e men a ion ime, 0 h, n = 2): Chao1 (78.50 ± 7.78), Shannon (3.53 ± 0.18),
Simpson (0.94 ± 0.02) and In e se Simpson (18.86 ± 7.63). These coe fi-
cien s showed a low amoun o a iabili y be ween ecal inocula and
eflec ed simila pa e ns in he mic obio a composi ion o homogena es.
Then, be a di e si y analysis o samples subjec ed o in i o ecal
e men a ion was pe o med o assess di e ences in mic obial di e si y
among subs a es (FOS and unmodified aga ose con ols, and β-D-F u-
(2→1)-D-Tag). Fo his pu pose, B ay-Cu is dissimila i y coe ficien s we e
calcula ed (Figu e S30) showing no significan di e ences (p> 0.05) among
subs a es. Be a di e si y dis ances we e also used o gene a e a dend og am
whe e samples co esponding o he same ecal pool inoculum clus e ed
oge he (Figu e S31). In addi ion, samples we e clus e ed acco ding o he
e men a ion ime (0, 8 o 24 h). These esul s highligh he e ec o in i o
e men a ion ime and a iabili y in he mic obio a composi ion a ibu ed
odi e en ecalinocula.β-D-F u-(2→1)-D-Tagand comme cialFOS we e
clus e ed in he same b anches e ealing simila modula o y p ope ies o
mic obial communi ies o hese wo subs a es compa ed o unmodified
aga ose con ol.
A P incipal Coo dina es Analysis (PCoA) (Figu e S32) was compu ed
o be e cha ac e ize sample dis ibu ion. Samples co esponding o each
ecal pool inoculum we e comple ely disc imina ed. In gene al, samples
we e also classified based on he e men a ion ime (0, 8 and 24 h). The
mic obio a composi ion ba plo e ealed cha ac e is ic bac e ial gene a
p ofiles o each subs a e (Figu e S33). In his ega d, ecal e men a ion o
comme cial FOS and β-D-F u-(2→1)-D-Tag led o high abundances o
Mi suokella and Bifidobac e ium while ecal e men a ion o unmodified
aga osecon olled oahighCollinsella abundance.These esul semphasize
di e ences in he mic obio a modula o y p ope ies be ween o iginal and
modified aga ose subs a es ha could be a ibu ed o hei s uc u al
ea u es.
To gain a be e unde s anding o mic obial axa changes induced by
each subs a e (FOS and unmodified aga ose con ols, and β-D-F u-
(2→1)-D-Tag), se e al di e en ial abundance me hods de eloped o
mic obiome s udies (ANCOM, LE Se, and me agenomeSeq) we e com-
pu ed. A o al o 25 gene a showed s a is ically significan inc emen s
(p < 0.05) in hei ead coun s and abundances a e 8 o 24 h o in i o ecal
e men a ion. Mic obial inc emen s in ead coun s we e classified as low
( aluesbe weenze oand he secondqua ile),mode a e( aluesbe ween he
second and he hi d qua iles), and high ( alues be ween he hi d and he
ou h qua iles) o compa a i e pu poses (Fig. 5).
Conce ning mic obial gene a showing low inc emen s, comme cial
FOS selec i ely p omo ed he g ow h o Anae o ib io, Co iobac e iaceae
UCG-003, In es inibac e , Megamonas and Tu icibac e (Fig. 5A). O iginal
aga ose con ol led o highe inc emen s o Alis ipes, Bu y icicoccus, Fla-
oni ac o , Pa abac e oides,andSenegalimassilia compa ed o β-D-F u-
(2→1)-D-Tag. In con as , β-D-F u-(2→1)-D-Tag selec i ely s imula ed
Egge hella, Ruminococcus gna us g oup, and Su e ella.
The h ee subs a es unde s udy led o mode a e inc emen s o
Acidaminococcus and En e ococcus while comme cial FOS and β-D-
F u-(2→1)-D-Tag p omo ed he g ow h o Rombou sia and Succini-
ib io (Fig. 5B). Simila ly, FOS con ol and β-D-F u-(2→1)-D-Tag led
o highe inc emen s o Bifidobac e ium and Mi suokella compa ed o
he o iginal aga ose con ol (Fig. 5C). In con as , o iginal aga ose
exe edahighe modula o y e ec onCollinsella and Megasphae a han
he es o subs a es (Fig. 5C). Finally, FOS con ol selec i ely s imu-
la ed he g ow h o Lac obacillus.Thesefindings unde sco e he simi-
la i ies in he e men a i e p ope ies o comme cial FOS mix u es and
β-D-F u-(2→1)-D-Tag.
Complemen a y co ela ion ne wo k analysis e ealed a posi i e
associa ion be ween Alis ipes and Pa abac e oides, wo gene a showing
highe inc emen s a e ecal e men a ion o aga ose con ol compa ed o
β-D-F u-(2→1)-D-Tag. Simila ly, nega i e associa ions be ween Bifido-
bac e ium, showing high inc emen s a e FOS and β-D-F u-(2→1)-D-Tag
e men a ion, and Collinsella, selec i ely p omo ed by o iginal aga ose
con ol, we e ound. Then, s a is ical associa ions be ween mic obio a
Table 1 | De e mina ion o suc ose, β-D- uc o u anosyl-
(2→1)-D- aga opy anose and β-D- uc o u anosyl-(2→6)-
β-D- uc o u anosyl-(2→1)-D- aga opy anose du ing in i o
small in es inal diges ion
Diges ion ime (h) Suc ose β-D-F u-
(2→1)-D-Tag
β-D-F u-(2→6)-
β-D-F u-(2→1)-
D-Tag
0 100.0 (3.5) 100.0 (17.5) 100.0 (0.5)
1 No de ec ed 110.6 (17.3) 88.3 (5.6)
2 No de ec ed 104.8 (11.8) 86.0 (4.6)
3 No de ec ed 99.8 (17.7) 78.0 (13.3)
Da a a e exp essed as he mean wi h he SD in b acke s (n=3) and ep esen he pe cen age o
ca bohyd a e de e mined agains i s ini ial concen a ion.
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 6
composi ion and lac a e and SCFAs le els a e ecal e men a ion we e
de e mined (Fig. 6and Figu e S34).
Lac a e and SCFAs de e mined in ecal e men a ion samples include
ace ic acid (45.22 ± 31.19 mM, ange: 0.56 –109.04 mM), bu y ic acid
(12.67 ± 14.48 mM, ange: 0.16 –54.46 mM), p opionic acid
(14.89 ± 12.52 mM, ange: 0.28 –42.83 mM), ale ic acid (1.51 ± 3.07 mM,
ange0.00 –11.08 mM) andlac a e(3.10 ± 2.53 mM, ange 0.00–8.16 mM)
(Figu e S34). Co ela ion analysis e ealed posi i e associa ions be ween
Alis ipes, Clos idium inocuuum g oup, Egge hella, Fla oni ac o , Pa a-
bac e oides andSenegalimassilia, and bu y ic and ale ic acids le els (Fig. 6).
Mos o hese axa and Succini ib io showed posi i e co ela ions wi h
p opionic acid le els. In addi ion, posi i e co ela ions be ween Bifido-
bac e ium, Lac obacillus,andEn e ococcus wi h ace icacidle elswe e ound
(Fig. 6). Simila ly, Bifidobac e ium, Lac obacillus, Mi suokella,andTu -
icibac e abundances we e posi i ely co ela ed wi h lac a e le els (Fig. 6). In
gene al,gene ashowingsignifican (p< 0.05)co ela ions wi h SCFAsle els
in ol e main SCFAs p oduce s om human gu mic obio a47.Inaddi ion,
posi i e associa ions be ween Lac obacillus, Tu icibac e ,andSuccini ib io
wi h se e al SCFAs including ace ic, p opanoic, and bu y ic acids ha e been
epo ed48, in ag eemen wi h ou esul s.
Se e al gene a p omo ed by hese subs a es we e also s imula ed by
o he unc ional ca bohyd a e ac ions in p e ious s udies. In his ega d,
in i o e men a ion o bagasse ac ions led o an inc emen in Collinsella,
Pa abac e oides,andSenegalimassilia while Bifidobac e ium was selec i ely
p omo ed by alkali-soluble a abinoxylo-oligosaccha ide ac ions49. Simi-
la ly, in i o e men a ion o apple pomace and pec in ac ions esul ed in
high abundances o Tu icibac e , a bac e ium wi h possible an i-
inflamma o y e ec s50,Lac obacillus and Succini ib io48.
I should be no ed ha β-D-F u-(2→1)-D-Tag exe s a highe bifi-
dogenic e ec han unmodified aga ose, simila o he one obse ed o
comme cial FOS mix u es. Bifidobac e ium species a e widely ecognized as
p obio ics ha exe nume ous benefi s on human heal h including he
Fig. 5 | S a is ically significan (adjus ed p- alues (padj) < 0.05) inc emen s in
bac e ial gene a a e ecal e men a ions o di e en subs a es: uc o-
oligosaccha ides (FOS con ol), unmodified aga ose con ols ( aga ose), and
uc osyla ed aga ose (β-D-F u-(2 →1)-D-Tag). These inc emen s we e classified
in o low (A), mode a e (B), and high (C) inc emen s compa ed o he ini ial e -
men a ion ime. Low inc emen alues we e comp ised be ween ze o and he second
qua ile. Simila ly, mode a e inc emen alues we e comp ised be ween he second
and he hi d qua iles. Finally, high inc emen alues we e comp ised be ween he
hi d and he ou h qua iles. T: e men a ion ime a which maximum inc emen o
a specific genus was obse ed. I: indi iduals (dono s) showing he maximum
inc emen o a genus in hei mic obio a (i should be no ed ha mos significan
di e ences we e obse ed in all indi iduals). P: abundance pe cen age o a specific
genus showing a maximum inc emen a a gi en ime. Bac e ial coun s inc emen s
a e shown in pa en heses.
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 7
p oduc ion o SCFAs51,52, in ag eemen wi hou esul s. These di e ences in
gu bac e ial selec i i y could be a ibu ed o he uc osyla ion o aga ose
ia he glycosidic linkage β-(2-1) which is associa ed wi h well-desc ibed
p ebio ic p ope ies. S uc u al modifica ion o aga ose also esul ed in a
highe capaci y o p omo e SCFAs-p oducing Mi suokella53 compa ed o
he o iginal aga ose.
The abili y o Bifidobac e ium species o use aga ose as subs a e has
been in es iga ed54. I has been epo ed ha aga ose e men a ion is
common among lac ic acid bac e ia, al hough se e al Bifidobac e ium may
no beable odeg ade aga ose55.Resul she ep esen edsugges ha aga ose
uc osyla ion may enhance i s e men a i e p ope ies leading o a highe
bifidogenic e ec .
In conclusion, SacB, a le ansuc ase om Bacillus sub ilis CECT 39,
specifically ans e s uc ose moie ies om suc ose o ei he C-1 o
D- aga ose o C-5 o L-so bose, o ming he main disaccha ides β-D-
uc o u anosyl-(2 →1)-D- aga opy anose and β-D- uc o u anosyl-
(2 →5)-L-so bopy anose, espec i ely.Molecula dockinghasp o ideda
s uc u al basis o unde s anding he ca aly ic mechanism o SacB and i s
p e e ence o hese a e suga s o e o he po en ial subs a es like
D-psicose. Addi ionally, he o ma ion o a e suga -based oligosaccha ides
wi h a highe deg ee o polyme iza ion was demons a ed h ough he
ans e o addi ional uc ose esidues o C-6 o ei he he β-2,1-linked
uc ose in D- aga ose o he β-2,5-linked uc ose in L-so bose, as de e -
mined by 2D-NMR. In i o s udies ha e shown he high esis ance o he
no el aga ose-based oligosaccha ides o he hyd oly ic ac ion o he
mucosal disaccha idases embedded in he pig small in es inal BBMV
h oughou he whole diges ion ime, hus suppo ing hei po en ial ole as
low calo ie ing edien s. The lack o diges ibili y o he biosyn hesized
aga ose-based ca bohyd a es wa an ed u he s udies dealing wi h hei
e men able p ope ies and po en ial abili y o selec i ely modula e he gu
mic obio a. Thus, β-D- uc o u anosyl-(2 →1)-D- aga opy anose exhib-
i ed dis inc modula o y p ope ies on he human ecal mic obio a com-
pa ed o unmodified aga ose, such as a highe bifidogenic e ec , simila o
ha obse ed o well-known p ebio ics like inulin- ype FOS. In e es ingly,
he uc osyla ion o aga ose h ough he glycosidic linkage β-(2 →1)
selec i ely s imula ed he g ow h o Egge hella,Ruminococcus gna us
Fig. 6 | Co ela ion hea maps showing he associa ions be ween sho -chain a y
acids (SFCAs) concen a ions (mM) and mic obial gene a exhibi ing s a is ically
significan inc emen s du ing ecal e men a ion o di e en subs a es: uc o-
oligosaccha ides (FOS con ol), unmodified aga ose con ols ( aga ose), and
uc osyla ed aga ose (β-D-F u-(2 →1)-D-Tag). Blue and ed do s indica e
posi i e and nega i e co ela ions exp essed as Pea son co ela ion coe ficien s.
Colo in ensi y is in p opo ion o magni ude. Da a ep esen ed co espond o all
e men a ion imes (8 and 24 h). SCFAs de e mined include ace ic, bu y ic, o mic,
iso ale ic, lac ic, p opanoic, and ale ic acids. To al SCFAs concen a ion is also
ep esen ed.
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 8
g oup, Rombou sia, Succini ib io,Su e ella o SCFAs-p oducing Mi suo-
kella as compa ed o he o iginal aga ose. The e o e, he biocon e sion o
D- aga ose in o β- uc osyla ed species wi h a highe deg ee o poly-
me iza ion is an e ficien app oach o inc ease he bioa ailabili y o aga ose
and p omo e i s in e ac ion wi h he gu mic obio a. The app oach
desc ibed in his wo k opens new oppo uni ies o ailo ing na u al a e
suga s, such as D- aga ose and L-so bose, o p oduce no el biosyn hesized
ca bohyd a es wi h unc ional and s uc u al p ope ies ha a e appealing
o hei use as eme ging p ebio ics and low-calo ic swee ene s.
Me hods
Chemicals, eagen s, and ca bohyd a es
Allchemicalsand eagen susedwe eo analy icalg adeand we e pu chased
om Sigma-Ald ich (S . Louis, MO, USA), VWR (Ba celona, Spain), and
Me ck (Da ms ad , Ge many). D-Taga ose was ob ained om Ca bosyn h
(Comp on, UK). Ul apu e wa e , p oduced in-house using a labo a o y
wa e pu ifica ion sys em (Milli-Q Syn hesis A10, Millipo e, Bille ica, MA,
U.S.A.), was used h oughou he expe imen s.
P oduc ion, pu ifica ion, and ac i i y assay o ecombinan
le ansuc ase SacB enzyme
Le ansuc ase SacB (EC 2.4.1.10) om Bacillus sub ilis CECT 39 (ATCC
6051) was o e p oduced in Esche ichia coli and pu ified ollowing he
me hod p e iously desc ibed by Díez-Municio, e al. 30. The o al ac i i y o
SacB was de e mined by measu ing he amoun o ee glucose eleased,
while he hyd oly ic ( uc osidase) ac i i y was assessed by quan i ying he
amoun o o med uc ose.
The ans uc osyla ion ac i i y ( ans e ed uc ose) was defined as
he di e ence be ween he amoun s o eleased glucose and uc ose.
Consequen ly, SacB exhibi ed a o al specific ac i i y o 20.77 uni s pe
millig am (U mg−1), whe e 1 uni is defined as he amoun o enzyme ha
eleases 1 μmolo glucosepe minu ea 37°Cinasuc oseconcen a iono
100 g L−1a pH6.0,usinga50 mMpo assiumphospha ebu e .Thespecific
uc osidase ac i i y was de e mined o be 10.79 U mg−1, whe e 1 uni is
defined as he amoun o enzyme ha eleases 1 μmolo uc ose pe minu e
unde he same condi ions. Finally, he ans uc osyla ion ac i i y was
calcula ed o be 9.98 U mg−1,defined as he amoun o enzyme equi ed o
ans e 1 μmol o uc ose pe minu e o o he molecules unde he es ed
condi ions32.
Enzyma ic syn hesis o a e suga s-based oligosaccha ides
The p oduc ion o aga ose-based ca bohyd a es, using suc ose as he
uc osyl dono and D- aga ose as he accep o , was ca ied ou h ough a
ans uc osyla ion eac ion ca alyzed by SacB. This eac ion ollowed he
condi ions (i.e., pH 6.0 in 50 mM po assium phospha e bu e a 37 °C)
p e iously op imized by Díez-Municio e al. 30 o syn hesizing lac ose-
based uc osyla ed oligosaccha ides. In his s udy, he SacB concen a ion
was op imized by es ing h ee di e en enzyme le els: 0.6, 3.1, and 6.2 U
( ans uc osyla ion ac i i y) o SacB pe mL o he eac ion mix u e.
Samples we e incuba ed in 2.0 mL indi idual ubes on an o bi al shake
(Eppendo The momixe Com o , Hauppauge, NY, U.S.A.) a 1000 pm.
The eac ions we e allowed o p oceed o up o 72 h, wi h aliquo s aken a
sui able ime in e als (0.5, 1, 2, 4, 6, 8, 24, 48, and 72 h). The enzyme was
inac i a ed by hea ing a 100 °C o 5 min, and he inac i a ed samples we e
hen s o ed a −20 °C un il analysis. Once he enzyme concen a ion was
op imized, he p oduc ion o aga ose-based oligosaccha ides was s udied a
a ious suc ose o D- aga ose concen a ion a ios, i.e., 200:300, 300:300,
200:450, 200:200 g L−1. The p ima y pa ame e s used o e alua e he e fi-
ciency o enzyma ic syn hesis o no el aga ose-based ca bohyd a es we e:
i) concen a ion (exp essed as g ams pe li e o eac ion); ii)yield
(exp essed as g ams pe 100 g o s a ing aga ose added); and iii)con e -
sion (%) (exp essed as he pe cen age o p oduc concen a ion pe con-
cen a ion o s a ing suc ose).
Likewise, addi ional eac ions based on D-suc ose:L-so bose and D-
suc ose:D-psicose we e ca ied ou unde he same op imized condi ions o
pH, empe a u e, and le ansuc ase concen a ion, and a a concen a ion
a io o 200:200 g L−1. All syn hesis eac ions we e pe o med in duplica e.
Pu ifica ion and isola ion o D- aga ose- and L-so bose-based
oligosaccha ides
Conside ing he lack o comme cially a ailable s anda ds o a e suga -
based oligosaccha ides, he main oligosaccha ides syn hesized a e 24 h o
he ans uc osyla ion eac ion based on D-suc ose:D- aga ose and D-
suc ose:L-so bose mix u es, ca alyzed by le ansuc ase unde he op imized
condi ionsdesc ibedea lie ,we eisola edandpu ifiedbasedon hei deg ee
o polyme iza ion. This was achie ed by p epa a i e LC-RID on an Agilen
Technologies 1260 Infini y LC Sys em (Boeblingen, Ge many), using a
Zo bax NH
2
P epHT p epa a i e column (250 × 21.2 mm, 7 μmpa icle
size) (Agilen Technologies, Mad id, Spain). Two mL o eac ion mix u es
(200 mg o o al ca bohyd a es) we e elu ed wi h ace oni ile:wa e (65:35,
: ) as he mobile phase a a flow a e o 21.0 mL/min o 30 min. The
sepa a ed di- and isaccha ides we e collec ed using an Agilen Technol-
ogies 1260 Infini y p epa a i e-scale ac ion collec o (Boeblingen, Ge -
many), and he ac ions we e pooled, e apo a ed in a o a y e apo a o
R-200 (Büchi Labo echnik AG, Flawil, Swi ze land) below 25 °C, and
eeze-d ied. Th ee isola ed ac ions, labeled as T1, T2, and T3, and ou
ac ions, labeled as S1, S2, S3, and S4, con aining he pu ified uc osyla ed
D- aga ose- and L-so bose-based oligosaccha ides, espec i ely, we e p e-
pa ed o subsequen NMR cha ac e iza ion. Likewise, pu ified T1 and T3
ac ions we e subjec ed o in i o gas oin es inal diges ibili y s udies, and
he pu ifiedT1 ac ion was also es ed o i s abili y o modula e he human
ecal mic obio a in i o.
S uc u al cha ac e iza ion o biosyn hesized a e suga s based-
oligosaccha ides by Nuclea Magne ic Resonance (NMR)
S uc u e elucida ion o he pu ified ac ions (T1-T3, S1-S4) was accom-
plished using Nuclea Magne ic Resonance spec oscopy (NMR). NMR
spec a we e eco ded a 298 K, wi h D2O as he sol en , on an Agilen
SYSTEM 500 NMR spec ome e (1H a 500 MHz, 13C a 125 MHz)
equipped wi h a 5 mm HCN cold p obe (25 K). Chemical shi s o 1H(δH)
inpa s pe million we e de e mined ela i e osol en esidual peak(HDO,
δH 4.79) and o 13C(δC) we e de e mined ela i e o ex e nal s anda d o
1,4-dioxane in D
2
O(δC 67.40). One-dimensional (1D) NMR expe imen s
(1Hand13C{1H}) we e pe o med using s anda d pulse sequences. 1H
expe imen s we e also acqui ed using PRESAT sequence o wa e peak
supp ession. Two-dimensional (2D) [1H, 1H] NMR expe imen s, including
g adien co ela ion spec oscopy (gCOSY) and o al co ela ion spec o-
scopy (TOCSY), we e ca ied ou wi h he ollowing pa ame e s: a delay
ime o 1 s, a spec al wid h o 2800 Hz in bo h dimensions, 2048 complex
poin s in 2, 4 ansien s o each o 128 (200 o TOCSY) ime inc emen s,
and linea p edic ion o 512. The da a we e ze o-filled o 2048 × 2048 eal
poin s. 2D [1H−13C] NMR expe imen s [g adien he e onuclea single-
quan um cohe ence (gHSQC), hyb id expe imen gHSQC-TOCSY, and
g adien he e onuclea mul iple bond co ela ion (gHMBC)] u ilized he
same 1H spec al window and a 13C spec al window o 7541.5 Hz, 1 s o
elaxa ion delay, 1024 da a poin s, and 128- o 200- ime inc emen s, wi h
linea p edic ion o256. The da a we eze o-filled o 2048 × 2048 ealpoin s.
Typical numbe s o ansien s pe inc emen we e 4 and 16. A mixing ime
o 80 ms was used o he gHSQC-TOCSY expe imen .
Quan i a ion o aga ose-based oligosaccha ides in enzyma ic
mix u es by Gas Ch oma og aphy-Flame Ioniza ion De ec o
(GC –FID)
The ca bohyd a es p esen in he eac ion mix u es and hei pu ified o ms
we e analyzed by GC-FID as ime hylsilyla ed oximes (TMSO), p epa ed
ollowing he me hod epo ed by B obs (1972)56. To o m ca bohyd a e
oximes,250 μLo hyd oxylaminechlo ide in py idine(2.5% w/ )was added
o d ied samples, and he mix u e was hea ed a 70 °C o 30 minu es. The
esul an oximes we e silyla ed using 250 μL o hexame hyldisilazane and
25 μLo ifluo oace ic acid a 50 °C o 30 min. The eac ion mix u es we e
h ps://doi.o g/10.1038/s41538-024-00320-8 A icle
npj Science o Food | (2024) 8:74 9
