Adjacent-possible ecological niche: growth of Lactobacillus species co-cultured with Escherichia coli in a synthetic minimal medium

In certain conditions, members of the Lactobacillus genus are auxotrophs that have fastidious requirements for growth. Notably, Lactobacillus cannot grow in M9 medium, a minimal synthetic medium used for Escherichia coli. However, we found that some Lactobacillus strains can be grown in M9 when co-cultured with E. coli K-12. In the co-culture, L. casei proliferates exponentially, reaching cell densities of 108 CFU (colony-forming unit) ml−1 in 6 h and dominating E. coli in the late growth phase. Spent medium from E. coli grown overnight lacked this growth-promoting effect on L. casei. Similarly, the effect was not observed when the species were separated by a 0.4-µm membrane. Microscopic observations showed that L. casei are embedded in the micro-scale clusters of E. coli in the early growth phase. This study describes for the first time the ability of a Lactobacillus species to grow in minimal medium when in close proximity with co-cultured bacteria.

NBRC 3831 with E. coli K-12, including the demonstration that E. coli fimbriae and lipopolysaccharide (LPS) are essential mediators of this interaction 28 .
In the present study, we report for the first time that E. coli K-12 supports the growth of some Lactobacillus species during co-culture in minimal medium via a process that requires cell-cell contact or close proximity, but does not require fimbriae or LPS. We consider the significance of this finding in terms of 'adjacent-possible ecological niches' in a microbial community.

Results
Eleven LAB from distinct sources, such as fermented foods, and two S. aureus strains, which are partly auxotrophic for amino acids and are close relatives of Lactococcus from a genome-wide viewpoint 29 , were tested for their growth in co-culture with E. coli K-12. Aside from K-12 itself, none of these strains could grow in mono-culture in M9 minimal medium (Fig. S2). We found that the CFU values of 6 of the tested Lactobacillus strains, including 3 belonging to the L. casei-group and 3 strains of L. plantarum, increased in the presence of E. coli (Fig. 1). In contrast, co-culture with E. coli did not increase the CFU values of L. fermentum, Leuconostoc mesenteroides, Pediococcus acidilactici, L. sakei, or Lactococcus lactis. Similarly, E. coli did not support the growth of the S. aureus strains. Thus, the growth support by E. coli was specific for LAB of the L. casei-group and L. plantarum.
Subsequent co-culture experiments focused on L. casei NBRC 3831 as a representative LAB. Mono-culture of L. casei in M9 showed little increase in CFU numbers ( Fig. 2A), whereas co-culture with E. coli supported growth in this medium, with co-cultured L. casei reaching a density of 10 8 CFU ml −1 after 6 h (Fig. 2C). During log-phase growth in the co-culture, the specific growth rate was 0.396 h −1 , approximately 67% that of the mono-culture in MRS (µ  0.6 h −1 ). After 24 h of co-culturing (Fig. 2B), L. casei dominated over E. coli.
Next, we monitored carbon consumption and metabolites of these two strains to evaluate metabolic activities during cell proliferation. The concentrations of glucose and of d-and l-lactate in the co-culture were measured (Fig. S3). During mono-culture in M9, L. casei consumed glucose and produced l-lactate (despite the lack of increase in CFU number) (Fig. S3A,B). These results showed that l. casei is metabolically active even under conditions that are insufficient for cell proliferation, which might provide a useful insight into persistency of health-promoting effects in terms of probiotics. During co-culture, l-lactate production was higher than in the mono-culture, probably due to the increase in CFU number of L. casei (Fig. S3B,C). E. coli, which is known to be a d-lactate producer, started to produce d-lactate after 10 h of mono-culture in M9; the same pattern was observed in the co-culture in this medium.
We next tested whether the spent medium or the cell-free extract obtained from mono-culture of E. coli in LB would support the growth of L. casei in the absence of co-culture (Fig. 3A,B). The spent medium from E. coli had no effect when provided at concentrations of 10 to 40% (percent volume equivalent) in M9 medium. These results demonstrated that the growth-supporting factor is not provided by the spent medium obtained from mono-culture of E. coli in LB. These observations also excluded the possibility that residual soluble factors carried over from the initial E. coli pre-culture supported L. casei growth during co-culture. In contrast, cell-free extract at 30% supported the growth of L. casei as well as that of L. sakei NBRC 3541 (Fig. 3C), a LAB species that was not able to grow when co-cultured with an E. coli in minimal medium.
Since the E. coli spent medium was insufficient to support the growth of L. casei in M9, we hypothesized that growth effects on L. casei require close proximity with E. coli cells. To test this theory, we used a membrane culture system wherein L. casei and E. coli were separated by a membrane (pore size, 0.4 µm). No growth support was observed for L. casei placed in M9 in a compartment adjacent to a parallel E. coli culture. Addition of L. casei to the E. coli compartment did not affect this result (Fig. 4, L+E vs. L). Thus, direct interaction with, or close proximity to, E. coli cells is required to permit the observed growth-supporting effect on L. casei growing in minimal medium (Fig. 4).
As noted above, we previously showed that fimbriae and LPS are essential mediators of some co-aggregation processes 28 . We postulated that the same might apply to the co-culture phenomenon described in the present work. Therefore, we evaluated the growth-supporting effect of E. coli using strains deleted for either fimA (the structural gene of the type I fimbriae 30 ) or rfaC (yielding a deep-rough LPS mutant with a truncated core LPS oligosaccharide 31 ). Notably, both of these E. coli mutants were able to support the growth of L. casei upon co-culture ( Fig. 5), demonstrating that the contact/proximity-based co-culture effect observed in the present work does not depend on fimbriae or LPS.
In the above co-culture experiments, the initial inoculum ratio was 10:1 (L. casei: E. coli). We wanted to test if the growth support could be observed even when L. casei existed as a minority component of the co-culture. As shown in Fig. 6, growth was robust even when the culture was initiated using L. casei at 10 4 CFU ml −1 , that is, at an initial inoculum ratio of 1:100 (L. casei: E. coli). This result suggested that the retrieval of the growth-supporting factor(s) from E. coli does not rely on having high initial density of LAB, in contrast to other systems, for instance, as in the induction of LAB bacteriocins by quorum sensing 32 .
We used Gram staining to observe cells after 3 h of co-culturing (Fig. 7A,B). The L. casei cells formed long chains. Individual L. casei cells were observed embedded in clusters of E. coli cells. To exclude a possible artifact during the fixation and Gram-staining procedure, an unfixed sample from 3-h co-culture of L. casei and E. coli was directly observed by fluorescence microscopy. In addition to the free E. coli cells (expressing GFP), clusters carrying embedded L. casei cells were observed (compare Fig. 7D). The viability of cells was analyzed by PI staining (Fig. 7C-E). PI is an intercalating DNA stain that penetrates injured membranes; thus, PI staining is indicative of dead cells. Notably, cells located within micro-scale clusters were often positive for PI staining (Fig. 7C,D) and were presumably dead. In E. coli mono-cultures, PI-positive cells were rare, though not undetectable (Figs 7E and S4). Considered together, these results indicate that L. casei has the ability to interact with E. coli cells and form micro-scale clusters accumulating PI-positive cells, which could explain how L. casei uptakes certain growth-supporting factors from E. coli during co-culture.

Discussion
This study demonstrated for the first time that LAB, which are strictly auxotrophic, can interact with co-cultured E. coli to proliferate in what would otherwise be nutrient-limiting conditions. Such growth ability was found to be specific for some Lactobacillus species, but the reason for the observed specificity remains to be clarified. This specificity is not simply attributable to the diversity of the nutrient auxotrophies among LAB species, because non-co-culturable species (such as L. sakei) were able to grow when cultured on minimal medium supplemented with an E. coli cell extract. It is also unlikely that the observed growth results from genetic reversion of auxotrophy by mutation 33 , given that a starting inoculum as small as 10 4 CFU ml −1 yielded a swift growth during co-culture. Secreted molecules such as antimicrobial peptides and chemoattractants would be among the candidates responsible for the observed strain specificity, however the spent medium of a mono-culture of L. casei NBRC 3831 in MRS did not induce PI-positive E. coli cells (Fig. S5). In this regard, it might be important to note that E. coli tended to assemble around L. casei NBRC 3831 in co-culture. For instance, CO 2 , necessary for growth of lactobacilli 34 , might be associated with the cell-cell contact or close proximity. It might be also important to note that growth simulation analysis of a co-culture of LAB and E. coli predicted that some E. coli-derived factors could improve LAB growth, although it was in a condition in which LAB can grow in the absence of E. coli 35 .
Macro-scale co-aggregation requires acidic conditions (below pH 5.0) 28 , whereas the pH in the exponential growth phase of the co-culture in this study was above 6.0 (Fig. 2D). This result suggested that the co-clusters observed in the present study are based on a mechanism distinct from the fimbria-or LPS-mediated mechanism used for macro-scale co-aggregation. Thus, we propose that the growth support observed concomitantly with the formation of micro-scale co-clusters relies on a mechanism distinct from the previously reported macro-scale co-aggregation.
It was not elucidated in the present study whether the micro-scale co-cluster itself is essential for the growth support. However, it is a new finding that L. casei can proliferate only when this LAB is provided with cell-cell contact or close proximity to E. coli, providing us a new ecological viewpoint of how an auxotrophic minority in a microbiota can exploit its ecological niche. Here, we would term such an ecological concept as an 'adjacent-possible ecological niche (APEN)' . The 'adjacent-possible' concept itself was previously proposed by the theoretical biologist Stuart Kauffman not only in the context of biology but also in a broad range of scientific fields, including economics 36,37 . Considering LAB as the minority (<0.1%) in the dense colonic microbiota (10 10 -10 12 cells/g) 38 as well as in the microbiome of fermented food, an ecological viewpoint of such an adjacency would have relevance in the context of LAB physiology. The better understanding of APEN may permit us to expand our scope beyond the one currently limited to a narrow set of interactions between LAB and yeasts 24 or cheese-isolated E. coli 35 and between L. bulgaricus and Streptococcus thermophilus 39 . The species specificities and overall relevance of APENs to the LAB lifestyle, as well as the bacterial strategies targeting APENs, are intriguing questions that will need be addressed in future studies.
In conclusion, L. casei group and L. plantarum strains became culturable in M9 medium when co-cultured with E. coli K-12. This growth support did not reflect "carry-over" of soluble nutrients from the pre-culture supernatant or the leakage of compounds from the mono-cultured E. coli cells. The growth of L. casei NBRC 3831 required cell-cell contact or close proximity to E. coli cells, suggesting that L. casei NBRC 3831 uptakes growth-supporting factor(s) from nearby E. coli cells.

Methods
Strains and culture conditions. The strains used in this study are listed in Table 1. The parent wild-type strain E. coli K-12 and the deletion strains ΔfimA (JW4277) and ΔrfaC (JW3596) of the KEIO collection were obtained from the National Institute of Genetics (Shizuoka, JAPAN). Lactobacillus strains were statically cultured in 10 ml of de Man, Rogosa, and Sharpe (MRS) medium (Oxoid, Hampshire, England) (10 g proteose peptone, 10 g beef extract, 5 g yeast extract, 20 g glucose, 1 g polysorbate, 2 g ammonium citrate, 5 g sodium acetate, 0.1 g  Construction of pRIT-Pldh-gfp and recombinant strains. The promoter region of the Listeria monocytogenes ldh (lactate dehydrogenase) gene was amplified with primers V25 (5′-TAAGTCGACTTTCTTGCCGTCCACAG-3′) and V19 (5′-CTAGTCGACTTCGAATTCCTCCTA-3′) from the genome of Listeria monocytogenes EGDe (http://genolist.pasteur.fr/ListiList/index.html). The resulting fragment was digested with SalI and then cloned into the SalI site of the pRIT-gfp plasmid 40 . The resulting plasmid, designated pRIT-Pldh-gfp, was transformed into E. coli K-12 BW25113 and JW4277 using a standard electroporation protocol 41 .
Co-culture conditions. Strains were individually pre-cultured as described above, and inoculated (at 1:100) alone or together into M9 medium (12.8 g Na 2 HPO 4 ·7H 2 O, 3 g KH 2 PO 4 , 0.5 g NaCl, and 1 g NH 4 Cl per liter of deionized water containing 2 mM MgSO 4 and 0.1 mM CaCl 2 ) supplemented with 2% glucose. For the time course analysis, 10 ml co-culture was statically cultured for 72 h at 37 °C. The necessity of cell-cell contact was tested in a 6-well Millicell ® plate (Millipore, Massachusetts, USA) with polyethylene terephthalate membranes (pore size, 0.4 µm) to prevent direct cell-cell contact between the two species. The total volume of a well consisted of 5 ml M9 medium, of which the upper and lower compartments held about 2 and 3 ml of the medium, respectively. The individual pre-cultures were inoculated at 1:100 into the respective compartments. The 6-well plate then was statically cultured at 37 °C. Growth was monitored by optical density (600 nm) and viable cell counts (CFU ml −1 ).
To determine the viable cell number, serial dilutions of the cultures were inoculated onto agar plates, and the CFUs (colony-forming units) were determined. CFU values are presented as the means ± standard errors of 3 independent experiments.
In order to determine the viable cell number from co-culture, we used LB agar medium for E. coli and pH modified MRS agar medium to specifically detect Lactobacillus colonies. The pH of the MRS agar was adjusted to 6.0 with acetate, which has an inhibitory effect on E. coli growth 42 . The CFU values of Lactobacillus species were comparable between normal MRS agar and acetate-MRS agar (Fig. S1A). The colonies on the acetate MRS agar were confirmed as Lactobacillus by sequencing the PCR-amplified 16 S rRNA genes with primers 341 F (5′-CCTACGGGAGGCAGCAG-3′) and 907 R (5′-CCGTCAATT CCTTT[A/G]AGTTT-3′) 43 (Fig. S1B).
Evaluation of the specific growth rate of L. casei NBRC 3831. The specific growth rate (μ) in the mid-exponential growth phase (3-6 h) of L. casei NBRC 3831 in MRS mono-culture and M9 co-culture was calculated by the following equation: where X t is the CFU value at hour t, and X t0 is the initial CFU value. Preparation of spent medium and cell-free extract from E. coli pre-culture. The cells and spent medium from a 10 ml E. coli pre-culture were separated by centrifugation at 1,500 g for 20 min at 4 °C. The supernatant was filter sterilized (0.2-µm cellulose membrane) and used for further analysis ("spent medium"). The collected cells were washed 3 times with phosphate-buffered saline (PBS, pH 7.0), re-suspended in 1.5 ml of PBS, and disrupted with sterile glass beads (diameter, 0.1 mm) using a cell disruptor (Disruptor Genie, Scientific Industries, NY, USA): a 2-ml tube containing 1.4 g beads and the cell suspension was mixed for 20 min with intervals to cool the tube in an ice-water bath. Following disruption, the volume of the mixture was raised to a total volume of 10 ml using PBS and centrifuged at 10,000 g for 20 min at 4 °C. The resulting supernatant was filtrated (0.2-µm cellulose membrane) and used as a cell-free extract for further analysis.

Strain Description Reference
Microscopic analysis. Gram staining was conducted using a commercial kit (Nissui Co., Tokyo, Japan); the stained samples were observed using a FSX100 inverted microscope (Olympus, Tokyo, Japan). Cell viability in a co-aggregate was evaluated in a 3-to 5-h co-culture of L. casei and GFP-expressing E. coli. The cells were stained with 3.5 µM propidium iodide (PI) 44 . Fluorescence microscopy was performed using a Leica TCS SP5 confocal laser microscopic system (Leica Microsystem Co., Wetzlar, Germany). The excitation wavelengths used were: 395 nm for GFP and 488 nm for PI. The emission was observed at 509 nm and 633 nm, respectively. Images were acquired and analyzed using Leica LAS AF software, version 2.6.0.
Determination of glucose and lactate concentrations. Glucose and lactic acids concentrations were measured using commercial kits according to the manufacturer's instructions. Glucose consumption was determined by the Glucose Test Kit Wako (Wako Chemical Co. Japan), and the production of l-(+)-and d-(−)-lactic acids was measured by a colorimetric method using the F-kit (Roche, Basel, Switzerland).

Statistical Analysis.
All experiments were conducted in triplicate. Values represent the mean ± SEM of three independent experiments. To compare two groups, student's test was used and P < 0.05 was considered as statistically significant.