Identification of genes encoding a novel ABC transporter in Lactobacillus delbrueckii for inulin polymers uptake

Lactobacillus delbrueckii JCM 1002T grows on highly polymerized inulin-type fructans as its sole carbon source. When it was grown on inulin, a > 10 kb long gene cluster inuABCDEF (Ldb1381-1386) encoding a plausible ABC transporter was suggested to be induced, since a transcriptome analysis revealed that the fourth gene inuD (Ldb1384) was up-regulated most prominently. Although Bacillus subtilis 168 is originally unable to utilize inulin, it became to grow on inulin upon heterologous expression of inuABCDEF. When freshly cultured cells of the recombinant B. subtilis were then densely suspended in buffer containing inulin polymers and incubated, inulin gradually disappeared from the buffer and accumulated in the cells without being degraded, whereas levan-type fructans did not disappear. The results imply that inuABCDEF might encode a novel ABC transporter in L. delbrueckii to “monopolize” inulin polymers selectively, thereby, providing a possible advantage in competition with other concomitant inulin-utilizing bacteria.

Inulin-type fructans (referred to, hereafter, as inulin) are common components of dietary fruits and vegetables, such as artichokes, chicory, bananas, garlic, and asparagus 1 . Inulin is a linear D-fructose polymer with β-(2-1) glycosidic bonds that is terminated by an α-(1-2)-glycosidic bond 2,3 -containing glucose molecule, in which the degrees of polymerization range from 2 to over 60 4 . Because of the β-(2-1)-linkages, inulin is not readily digested by humans but can be degraded and utilized by some intestinal bacteria with probiotic properties, such as lactobacilli and bifidobacterial. Inulin has thus been recognized as a prebiotic that can selectively stimulate probiotic bacterial growth, activity, or both, in the colon, thereby improving the host's health.
Probiotic bacteria are known to degrade inulin with extracellular enzymes, collectively termed inulinase or β-fructofuranosidase (SacA) 5 , into shorter fractions that include sucrose, fructose, and inulin with a degree of polymerization (DP) up to 6. These fractions are then taken up by the bacterial cells via membrane-bound proteins, known as sugar transporters, such as phosphotransferase system (PTS) and ABC transporter [6][7][8][9][10][11][12] . Tsujikawa et al. 13 conducted an experiment in which Lactobacillus paracasei DSM 20,020 and L. delbrueckii JCM 1002 T were grown in the presence of inulin (DP of > 8) and found that the former degraded the inulin, releasing free fructose into the media during growth, which was clearly detectable by thin layer chromatography (TLC), whereas the latter did not show such extracellular fructose accumulation. The evidence suggested that JCM 1002 T did not degrade inulin extracellularly but transported it whole into its cells for further metabolism. The bacteria Sphingomonas sp. strain A1 has been reported to transport alginate, another highly polymerized carbohydrate, into their cells through the ABC transporter system 14 . Therefore, we were compelled to investigate if the L. delbrueckii is equipped with a similar macromolecule-transport system for inulin.
This study aimed to identify the genes responsible for the transport of highly polymerized inulin (DP of > 8), since this identification leads to the discovery of a novel sugar transporter. To this end, global gene transcription profiles were employed to nominate candidate genes involved in inulin transport. A wide region of DNA containing the transcriptional units of the most upregulated candidate genes in the presence of inulin was then introduced into the amyE region of Bacillus subtilis 168, which does not grow when inulin is the sole carbohydrate source. Any transformant that grew with inulin was densely suspended in PBS containing inulin and incubated. During the incubation, we monitored the disappearance of the inulin fractions from the PBS and Selection of candidate genes involved in inulin transport by RNA-seq. JCM 1002 T was grown either in the presence of glucose or inulin for 6 h, and we performed the RNA-seq analysis to compare the transcriptome profiles. We selected genes belonging to the following two criteria stepwise; the first criterion consists of genes induced to more than double after 6 h of growth in the presence of inulin, and among them, the second criterion consists of genes repressed to less than half in the presence of glucose ( Supplementary Fig. 1). We expected that this selection would allow us to narrow down genes induced by inulin specifically, but in fact, there were too many such genes for this to be possible. The difference in the transcriptome indicated that the gene sets involved in or affecting the metabolism of glucose and inulin are completely different genetic entities. We thus used cellobiose as an alternative carbon source, which is composed of 2 molecules of D-glucose linked   Fig. 2). Subsequently, we compared the transcriptome profiles of JCM 1002 T cells grown either with cellobiose or inulin, indicating that the expression levels of the majority of genes did not change significantly ( Supplementary  Fig. 3). We analyzed a total of 1,587 genes to determine their FPKM value. As a result, only 51 genes satisfied the first criterion and the second criterion with cellobiose instead of glucose. We then screened the 51 genes with the third criterion where the genes tentatively encoding a transporter or whose function is unknown, and nominated a total of 22 candidate genes which might include genes responsible for the inulin transporting as listed in Supplementary Table 1. The increase in gene expression levels in inulin compared to cellobiose was between 2.1-fold and 3.5-fold.
To validate the expression data obtained from the RNA-seq experiments, the transcript levels of the 22 candidate genes were measured by RT-PCR (Table 1). The expression of 16 genes was reproducibly induced in the presence of inulin compared with cellobiose (between 1.0-fold and 4.4-fold). Among them, Ldb0442 (3.56-fold) and Ldb1384 (4.44-fold) showed the highest expression increase and were the most interesting candidate genes.

Identification of genes involved in inulin transport.
To examine whether the two candidate genes are responsible for the inulin transporter, Bacillus subtilis 168, which, like L. delbrueckii, belongs to Firmicutes and is easy to genetically manipulate, was utilized. Conveniently for this validation, B. subtilis 168 is incapable of growing in the presence of inulin (Fig. 2b 1-6). To investigate the functions of the candidate genes, wide regions containing the transcriptional units for Ldb0442 (Ldb0438-0448) and Ldb1384 (Ldb1381-1386), including adjacent intergenic regions, were introduced into the amyE region of B. subtilis 168 (designated UT01 and UT02 strains, respectively) to test if the exogenously introduced gene clusters would allow B. subtilis 168 to utilize inulin (Fig. 2).
Six clones of UT01 (amyE::Ldb0438-0448) and UT02 (amyE::Ldb1381-1386), together with the 168 strain, were cultured on glucose or inulin as a sole carbon source. All clones grew well in the presence of glucose as expected (Fig. 2a). Although no growth was observed for any clone of the parental 168 strain or UT01, all clones of UT02 grew well in the presence of inulin, indicating that the Ldb1381-1386 region is responsible for inulin unitization, probably involving inulin transport, and thus the region was designated inuABCDEF (Fig. 2b).
To confirm that the inuABCDEF introduced with a chloramphenicol cassette (cat), but no other region(s), in the UT02 strain was responsible for the inulin-utilization phenotype, a back-cross experiment was performed by introducing the inuABCDEF with the cat, which was amplified from the UT02 strain ( Supplementary Fig. 4). As a result of comparing colony formation on minimal medium agar plates containing chloramphenicol, inulin, or both, almost the same number of transformants appeared under all conditions, proving that those that can www.nature.com/scientificreports/ use inulin always acquire chloramphenicol resistance as well, and that inulin can be used as long as inuABCDEF is acquired.

Inulin is transported by an ABC transporter encoded by inuABCDEF in B. subtilis cells. To
investigate inulin transport into the UT02 cells, B subtilis cells with (UT02) or without (168 strain) an ABC transporter homolog encoded by inuABCDEF were incubated in PBS with inulin, and the extracellular and intracellular inulin levels were analyzed using HPLC. In the case of the extracellular inulin of 168-strain cells, the HPLC pattern of initially observed multiple peaks corresponding to inulin polymers did not change, even 2 h after the start of the incubation; whereas, for the UT02 cells, the peaks began to shrink after 0.5 h and disappeared after 2 h (Fig. 3). In contrast, in the case of the intracellular inulin of 168-strain cells, no peak corresponding to inulin emerged, even after 2 h, while for the UT02 cells, inulin peaks were clearly present after 1 h and increased after 2 h (Fig. 4). These results indicated that inuABCDEF encoding the ABC transport homolog in the UT02 stain was responsible for inulin transport.

Substrate specificity of the ABC transporter encoded by inuABCDEF.
To investigate the substrate specificity of the transporter encoded by inuABCDEF, the ability of the UT02 strain cells to take up other oligosaccharides (isomalto-oligosaccharide: IMO and xylo-oligosaccharide: XOS) and polysaccharides (alginate and levan) was examined by analyzing extracellular levels using TLC. As predicted, the amount of extracellular inulin remained unchanged after B. subtilis 168 cells were incubated for 2 h (Fig. 5a lane 1), while that decreased after UT02 cells were incubated for 2 h (Fig. 5a lane 2), On the other hand, there was no change to the amount of extracellular IMO after UT02 cells or the B. subtilis 168 cells were incubated for 2 h (Fig. 5b). The same applied to the XOS (Fig. 5c), alginate (Fig. 5d), and levan (Fig. 5e). These results indicated that inuABCDEF is not involved in the transport of IMO, XOS, alginate, or levan but is specifically associated with inulin transport.

Introduction of the inuABCDEF into 168 mutant strains. A homology search showed that inuAB-
CDEF only encodes an ABC transporter and not any genes involved in inulin degradation, even though the inuABCDEF-introduced B. subtilis strain, UT02, grew in the presence of inulin. This fact implies that the intracellular inulin transported by Ldb1381-1386 was utilized after degradation by an enzyme that B. subtilis 168 originally possessed. SacC has been reported to be a levanase that is not only capable of hydrolyzing levan but also inulin and sucrose 15 . According to UniProt, SacA and LevB belong to the same glycosyl hydrolase 32 family as SacC. In the case of LevB, endo-activity specifically hydrolyzing (β-2,6) fructosyl bonds has been reported 16 . Thus, to examine if these enzymes were responsible for the inulin degradation in the UT02 strain, sacA, levB, and sacC genes were deleted in the UT02 strain, and the resultant clones were designated UT03, UT04, and UT05, respectively (Supplementary Fig. 5). UT03 (△sacA) and UT04 (△levB) were seen to grow in the presence of inulin, while UT05 (△sacC) showed no growth ( Supplementary Fig. 6). Consistent with this phenotype, amount of the intracellular inulin polymers in the UT05 cells were increased compared to that in the UT02, indicating that the intracellular inulin in the UT05 cells was not digested in the absence of SacC as done in the UT02 cells (Fig. 6). www.nature.com/scientificreports/ nominate candidate genes involved in a novel inulin transporter by differential gene transcription 10,11 between glucose-grown cells and inulin-grown cells. However, expression levels of enormous genes showed a significant difference, and we failed to narrow down the candidate genes induced by inulin specifically. This failure was www.nature.com/scientificreports/ probably due to carbon catabolite repression (CCR) 17 . CCR is one of the most fundamental and highly conserved mechanisms in bacteria to repress various genes in the presence of a preferred carbon source such as glucose 18 . JCM 1002 T belongs to low-GC Gram-positive bacteria. Thus, its CCR may depend on catabolite control protein A (CcpA) as the principal DNA-binding regulator. CcpA forms a complex with P-ser-HPr to bind to its target sites with the consensus sequence named cre to repress the genes under CCR 19 . We found that most of the genes repressed in the presence of glucose were associated cre sequences. However, for a proper discussion of CCR, the results obtained in this study are still preliminary, and the further discussion of CCR remains for another time in the future. As a result of RT-PCR, Ldb0442 and Ldb1384 were selected as the leading candidate form the 22 candidate genes. And then, by use of a complementation test using B. subtilis, which is originally not able to utilize inulin, as a host, a gene cluster containing the Ldb1384 was finally found to be a candidate for an inulin transporter, since the cluster-introduced strain became to utilize an inulin as a sole carbon source. Although for some Lactobacillus, genetics to identify genes involved in utilization of inulin-type fructans from candidate genes exist, such as plasmid integration via homologous recombination 10 and Cre-lox-based mutagenesis system 11 , however, such methods could not be applied to the JCM 1002 T so far. Thus, this system to use the B. subtilis as a host to test function of genes would be useful not only for JCM 1002 T but also for some bacteria whose genetics has not been established.
In this study, we demonstrated that the gene cluster Ldb1381-1386, designated inuABCDEF, encodes a novel inulin transporter of JCM 1002 T required for the uptake of highly polymerized inulin-type fructans (DP of > 8).
There are some reports of bacterial strains that can directly transport short inulin-type fructans into their cells. L. plantarum WCFS1 can take up 1-kestose (DP 3) and nystose (DP 4) through a sucrose-transport system 20 . Some Bifidobacterium strains can transport oligofructose (DP 2 to 8) directly into their cells [21][22][23] . However, with regards to transporters that directly take up highly polymerized inulin-type fructans (DP of > 8), Widodo et al. 24,25 reported that Lactobacillus casei AP takes up inulin via the PTS or ABC transporter, degrading it intracellularly. However, the inulin-transport-associated genes have not been identified. Therefore, as far as we know, this is the first report on the identification of a gene cluster responsible for the transport of highly polymerized inulin-type fructans (DP of > 8) into bacterial cells. According to GenBank, inuAB (Ldb1381-1382), inuCD (Ldb1383-1384), and inuEF (Ldb1385-1386) are annotated to encode oligopeptide ABC transporter substrate binding protein (accession no. CAI98182 and CAI98183), oligopeptide ABC transporter permease (accession no. CAI98184 and CAI98185), and oligopeptide ABC transporter ATP-binding protein (accession no. CAI98186 and CAI98187), respectively. Therefore, inuABCDEF may also have "oligopeptide ACB transporter" function in addition to inulin transporter function. Reportedly, putative oligopeptide ATP transporters are responsible for sugar and oligosaccharide transport in Thermotoga maritima 26 . Although, as far as we know, there is no evidence that the oligopeptide ABC transporter functions as a polysaccharide transporter. To investigate the substrate specificity of the transporter encoded by inuABCDEF, we introduced inuABCDEF into the 168 strain, and the bacteria's ability to take up oligosaccharides (IMO and XOS) and alginate were investigated by TLC. The results suggested that inuABCDEF does not encode the transporter for these sugars. IMO is a mixture of glucose oligomers with α-(1-6)-linkages, while XOS is a mixture of xylose with β-(1-4)-linkages. Alginate is composed of blocks of mannuronic acid residues, blocks of guluronic acid residues, and blocks of alternating M and G residues. Moreover, our results show that levan does not seem to be recognized by inuABCDEF. Inulin is a D-fructose polymer linked by β-(2-1)-glycosidic bonds with a terminal glucose moiety that is linked by an α-(1-2)-glycosidic bond, as in sucrose. Correspondingly, levan possesses β-(2-6)-linkages as its skeleton, with branching β-(2-1)-linkages 27 . Therefore, the ABC transporter encoded by inuABCDEF may recognize β-(2,1)-fructosyl-linkages or 1-alpha-D-glucopyranosyl-2-beta-D-fructofuranoside. A future investigation of the specific substrate of the transporter will be necessary to elucidate the function of this novel transporter in more detail.
This study showed that JCM 1002 T has a dedicated transportation system for taking up whole inulin polymers. Moreover, as shown in our last report, JCM 1002 T grows better on inulin-type fructans than on fructose 13 . There are some cases in which bacteria grow better when fed oligosaccharides than their degradation products. For example, Bifidobacterium animalis DN-173 010 grows on inulin-type fructans Raftilose Synergy1 and Raftilose www.nature.com/scientificreports/ P95 (powders containing inulin and enzymatically digested inulin, respectively) but not on fructose. This suggests the adaptation of the strain to oligosaccharide metabolism provides advantages when competing with other microorganisms in the human gut, where oligosaccharides are the main energy resource 28 . L. delbrueckii is derived from plants, and inulin exists in a variety of plants, such as some fruit, artichokes, chicory, garlic, and asparagus in natura 29 . Plants are internally inhabited by diverse microbial communities, including bacterial, archaeal, fungal, and protistic taxa 30 , which represent the many organisms that L. delbrueckii needs to compete with for survival in the ecosystem. For this reason, in plants where inulin is present, L. delbrueckii may have evolved the ability to transport whole inulin polymers directly into their cells to out-compete other concomitant inulin-utilizing bacteria, as the extracellular hydrolysis of inulin provides monosaccharides for opportunistic competitors. As far as we have been able to determine, there is no gene cluster in any other species that could be considered identical to inuABCDEF based on sequence homology. In addition to Bifidobacterium animalis DN-173 010, other bacteria that demonstrate unique sugar transportation according to each environment have been frequently reported. For example, human milk oligosaccharide-specific ABC transporters of Bifidobacterium help in the adaptation and dominance of the bacteria in the infant gut ecosystem 31 . L. plantarum isolated from humans, unlike the strains isolated from fermented food, lack the genes for FOS metabolism, indicative of their evolution and adaptation to different environmental niches 32 . In their probiotic roles, L. delbrueckii are reportedly responsible for enhanced host immune responses and have antioxidative activity [33][34][35] ; therefore, L. delbrueckii promotes GI health in humans and other animals. However, some authors have reported that L. delbrueckii is rarely found among the gut microbiota after ingestion because of its limited capability to survive the restrictive conditions of GI digestion 36,37 . Other researchers have provided evidence that some L. delbrueckii strains can survive GI transit, but the number of bacteria decreases considerably in those cases 38 . If any of the probiotic strains of L. delbrueckii have a dedicated transportation system for the more highly polymerized inulin-type fructans, it may be possible to use such sugars as prebiotic supplements to selectively stimulate their growth in the host's digestive tract, especially as host cells have difficulty absorbing and utilizing these fructans.
In conclusion, by selecting 22 candidate genes with RNA-seq and RT-PCR and identifying those involved in inulin utilization by introducing them into the amyE region of B. subtilis 168, we showed that inuABCDEF encodes a novel inulin transporter that directly transports inulin into the JCM 1002 T cells. This novel inulin transporter may form part of the system through which L. delbrueckii gains an ecological advantage in bacterial communities: those bacteria only degrading inulin extracellularly may, although temporally, release sizeable amounts of fructose molecules, which are then readily "scavenged" by other bacteria that only utilize fructose; whereas strains, like L. delbrueckii, that take up whole inulin for intracellular digestion can "monopolize" inulin resources.

Methods
Bacterial strains, media, and substrates. The bacterial strains used in this study were L. delbrueckii JCM 1002 T , Bacillus subtilis YK05, and B. subtilis 168. JCM 1002 T was stored at − 80 °C in de Man-Rogosa-Sharpe (MRS) broth (Oxoid, Basingstoke, United Kingdom), and Bacillus strains were stored at − 80 °C in LB broth (Oxoid, Basingstoke, United Kingdom) until use. mMRS 39 was used as the basal fermentation medium throughout this study. The pH of the medium was adjusted to 6.5 before sterilization (121 °C for 15 min). Glucose, cellobiose, or inulin (Wako Pure Chemical Industries) was added to mMRS as a carbon source. The DP of inulin (Wako Pure Chemical Industries) is reported to vary between 3 and 60 according to information supplied by the company. In all cases, these sugars were sterilized by membrane filtration using Millex Syringe Filter Units (pore size, 0.45 µm) and added aseptically to the sterile mMRS medium. Minimal medium was used to investigate the utilization of sugar by transformants. Minimal medium consisted of 1% glucose, 0.5% Spizizen's minimal salts 40  Fermentation experiments. JCM 1002 T was cultured anaerobically in MRS broth at 37 °C for 12 h, after which the cultures were centrifuged at 15,000 rpm for 5 min. The bacterial pellet was then washed once with phosphate-buffered saline (PBS, 0.8% NaCl, 0.02% KH 2 PO 4 , 0.115% Na 2 HPO 4 pH 7.4) and re-suspended in PBS until the optical density at 660 nm (OD660) reached 0.4, followed by spotting 50 µL of this suspension onto 5 mL of mMRS containing 2% glucose (wt/vol), 2% cellobiose (wt/vol), or 2% inulin (wt/vol). Incubations of the media were performed anaerobically at 37 °C for up to 48 h, during which their OD660 were measured at 3, 6, 9, 12, 24, and 48 h after incubation. Anaerobic conditions were provided using AnaeroPack-Anaero (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan).

RNA-seq.
JCM 1002 T cells were cultured in 5 mL of mMRS containing 1% glucose (wt/vol), 1% cellobiose (wt/vol) or 1% inulin (wt/vol) anaerobically at 37 °C for 6 h. After centrifugation at 15,000 rpm for 5 min, the bacterial cells were harvested, and total RNA was extracted and purified using the Bacterial RNA kit (OMEGA) according to the protocol provided by the manufacturer. Quality control of each RNA sample was performed with an Agilent 2100 Bioanalyzer (Agilent Technologies, USA). The cDNA libraries were constructed using NEBNextUltra RNA library Prep Kit (NEB, USA) and submitted for sequencing using Illumina Hiseq 4000 (Illumina). Library construction and sequencing were performed by Allwegene Biotechnology Co., Ltd. ( www.nature.com/scientificreports/ The differentially expressed genes between glucose and inulin and between cellobiose and inulin were identified considering both fold changes, respectively. In RNA-seq expression data, the genes that satisfied the following three criteria were selected as candidate genes for an inulin transporter. www.nature.com/scientificreports/ cassette, and the 3' end of amyE were then assembled using the Gibson Assembly Master Mix (New England Biolabs E2611S) according to the manufacturer's instructions. The primers used are listed in Supplementary Table 3. The assembled PCR fragments were then introduced into the amyE region by transformation 42 , and transformants were selected on minimal medium containing 1% glucose (wt/vol) and chloramphenicol (5 µg/ml) at 37 °C for 48 h. Six clones were then transferred to minimal agar media containing 1% glucose (wt/vol) or 1% inulin (wt/vol) and incubated at 37 °C for 48 h.
Back-cross experiment. The Ldb1381-Ldb1386 region of the UT02 strain was amplified by a primer set for Ldb1380F/Ldb1387R, and the PCR product was introduced into B. subtilis 168 as described above. Transformants were selected on minimal medium containing chloramphenicol (5 µg/ml), 1% inulin (wt/vol), or both, at 37 °C for 48 h.
Identification of genes involved in inulin degradation in B. subtilis. Deletion mutants of sacC, levB, and sacA were obtained by substitution with the erythromycin resistance cassette from the BKE 43 collection in the National BioResource Project. The deletions were introduced into the UT02 strain and selected on minimal medium containing erythromycin (1 µg/ml) and lincomycin (12.5 µg/ml) at 37 °C for 48 h. Then, transformants were inoculated into minimal medium containing 1% glucose (wt/vol) or 1% inulin (wt/vol), and growth was observed after 48 h incubation at 37 °C.

Data availability
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.