Bioprocess Biosyst Eng (2016) 39:1527–1537
DOI 10.1007/s00449-016-1628-3
ORIGINAL PAPER
A newly isolated and identified vitamin B12 producing strain:
Sinorhizobium meliloti 320
Huina Dong1,2 • Sha Li1 • Huan Fang1,2 • Miaomiao Xia1,2 • Ping Zheng1,2
Dawei Zhang1,2 • Jibin Sun1,2
Received: 30 March 2016 / Accepted: 16 May 2016 / Published online: 9 June 2016
Ó Springer-Verlag Berlin Heidelberg 2016
Abstract Vitamin B12 (Cobalamin, VB12) has several
physiological functions and is widely used in pharmaceu-
tical and food industries. A new unicellular species was
extracted from China farmland, and the strain could pro-
duce VB12 which was identified by HPLC and HPLC–MS/
MS. 16S rDNA analysis reveals this strain belongs to the
species Sinorhizobium meliloti and we named it S. meliloti
320. Its whole genome information indicates that this strain
has a complete VB12 synthetic pathway, which paves the
way for further metabolic engineering studies. The optimal
carbon and nitrogen sources are sucrose and corn steep
liquor (CSL) plus peptone. The optimal combination of
sucrose and CSL was obtained by response surface
methodology as they are the most suitable carbon and
nitrogen sources, respectively. This strain could produce
140 ± 4.2 mg L-1 vitamin B12 after incubating for 7 days
in the optimal medium.
Keywords Vitamin B12  Sinorhizobium meliloti 
Cobalamin  Whole genome sequence  Response surface
methodology
H. Dong, S. Li are contributed equally to this work.
& Ping Zheng
zheng_p@tib.cas.cn
& Dawei Zhang
zhang_dw@tib.cas.cn
1 oxygen-dependent pathway exists in Rhodobacter sphaer-
Tianjin Institute of Industrial Biotechnology, Chinese
Academy of Sciences, Tianjin 300308, China
2 independent pathway has been discovered in Bacillus
Key Laboratory of Systems Microbial Biotechnology,
Chinese Academy of Sciences, Tianjin 300308, People’s
Republic of China
•
Introduction
Vitamin B12 (Cobalamin, VB12) is one of the largest
known non-polymeric natural compounds, as well as one of
the most studied compounds [1]. It is involved in wide
range of biochemical processes, such as DNA synthesis
and regulation, fatty acid synthesis, amino acid metabolism
as well as energy production [2]. At present, VB12 has been
widely used for the treatment of pernicious anemia and
peripheral neuritis [3], and it was found that VB12 was the
essential cofactor for methionine synthase and (R)-
methylmalonyl-CoA mutase in animals and humans [4].
VB12 is composed of a central cobalt atom that is
coordinated by a tetrapyrrole framework termed the corrin
ring and by two axial ligands. In the biologically active
forms of cobalamin, adenosylcobalamin (coenzyme B12,
Ado-Cbl) and methylcobalamin (Me-Cbl), the upper ligand
consists of either 5-deoxyadenosine or a methyl group,
respectively (Fig. 1), and the lower ligand, usually 5,6-
dimethylbenzimidazole (DMBI), is attached to the corrin
ring via the nucleotide loop. The final products occurring in
nature from vitamin B12-biosynthesis are Ado-Cbl and Me-
Cbl, while vitamin B12 is by definition cyanocobalamin
(CN-Cbl), which represents the form mainly manufactured
by industry [5].
The production of VB12 has received great attention due
to the increasing global needs. The microbial synthesis of
VB12 takes place via two different pathways, an oxygen-
dependent pathway and an oxygen-independent pathway,
each of which involves more than 20 enzymatic steps. The
oides and Pseudomonas denitrificans, while the oxygen-
megaterium, Salmonella enterica, and Propionibacterium
freudenreichii [6]. Propionibacterium denitrificans and P.
123
1528
Fig. 1 Structure of vitamin B12. R stands for different ligands of
vitamin B12: 5-deoxyadenosine (adenosylcobalamin), CH3 (methyl-
cobalamin), CN (cyanocobalamin), OH (hydroxylcobalamin)
freudenreichii are used exclusively for the industrial pro-
duction of VB12 [5]. The highest titers of VB12 by these
two strains have been reported to be 214 and 206 mg L-1,
respectively [6]. Propionibacterium freudenreichii belongs
to the generally recognized as safe (GRAS) category of
microorganisms [7], but its use for the production of VB12
for human consumption is complicated. VB12 production
by P. freudenreichii is performed via a two-step process,
where pseudo-cobalamin is first produced under anaerobic
conditions and DMBI-containing coenzyme B12 is then
produced from the pseudo-cobalamin under (partially)
aerobic conditions [8]. On the contrary, P. denitrificans
grows and synthesizes the entire structure of coenzyme B12
via a one stage process aerobically [9]. However, P. den-
itrificans grows slowly to a low cell density, has compli-
cated and less understood carbon metabolism, and not well
established for genetic toolbox for strain improvement
[10].
Sinorhizobium meliloti is a Gram-negative a-pro-
teobacterium that exists either as a free-living member of
the rhizosphere or as an intracellular nitrogen-fixing sym-
biont of legumes of the genera Medicago, Trigonella, and
Melilotus [11]. Cobalamin has been isolated from S.
meliloti and shown to contain DMBI as the lower ligand
[12]. The S. meliloti genome contains a complete set of
genes involved in aerobic cobalamin biosynthetic pathway.
123
Bioprocess Biosyst Eng (2016) 39:1527–1537
Thus, it is one of many bacteria capable of synthesizing
cobalamin de novo through an oxygen-dependent pathway
[1]. Sinorhizobium meliloti may be a candidate to replace
P. freudenreichii or P. denitrificans as industrial VB12
producer in the future.
In this study, a newly S. meliloti stain was isolated from
soil and was named S. meliloti 320. The product VB12 was
identified by HPLC and electrospray ionization (ESI) mass
spectrometry HPLC–MS/MS. Then, the strain was identi-
fied by sequence 16S rDNA and the genome. The fer-
mentation medium and culture conditions were optimized
to maximize the production of VB12. The results showed
that this strain can produce 140 mg L-1 VB12, which lay
the foundation for it to be used as an industrial VB12
producer.
Materials and methods
Species sampling and pre-cultures
Sinorhizobium meliloti 320 was isolated from private
farmland, Tianjin, China and the owner of the land gave
permission to conduct the study on this site. In this study, S.
meliloti 320 was identified as one out of 50 isolated bac-
terial strains due to its growth characteristics and its red
color [13]. Sinorhizobium meliloti 320 was purified by
serial dilutions and plating on 1.5 % minimal medium agar
containing: KH2PO4, 2.5 g L-1; NaCl, 2.5 g L-1; NH4Cl,
0.5 g L-1; MgSO4, 0.13 g L-1; Glucose, 1 g L-1 and
CaCl2, 5 g L-1. pH values of media were titrated to pH 7.0
with 1 M HCl. Plates were incubated for 1–2 days at
30 °C. After colony formation single colonies were iso-
lated and cultivated in 250 mL-flasks. Fermentation media
contained sucrose, 50 g L-1; CSL, 60 g L-1; (NH4)2SO4,
2 g L-1; CoCl2, 0.006 g L-1; KH2PO4, 2.6 g L-1; betaine
(N,N,N-trimethylammonioacetat), 6 g L-1 and DMBI
1 g L-1 at pH 6.8–7.0 (solid KOH, adjusted before auto-
claving). Individual colonies were incubated at 30 °C in
fermentation flasks at 200 rpm for 4 days.
Extraction and analysis of VB12: mass spectrometry
and quantification
VB12 concentration in fermentation broth was determined
by high performance liquid chromatography (HPLC) with
the method reported previously [14]. Broth sample (4 mL),
to which 1 mL of NaNO2 8 % (w/v) and 1 mL of glacial
acetic acid were added, was boiled for 30 min. The mixture
was then centrifuged at 10,000 rpm for 1 min and the
supernatants was filtrated through a 0.22-lm membrane
([ = 0.22 lm) filter. Then, 20 lL of NaCN 2 % (w/v)
was added to 1 mL of the supernatants. To separate the
Bioprocess Biosyst Eng (2016) 39:1527–1537
compounds aliquots (20 lL) were injected and analyzed
using an Agilent 1260 HPLC with a 5C18-250A column
(Agilent, 4.6 mm id 9 250 mm, 5 lm) thermostated at
35 °C. The mobile phase consists of water: acetonitrile (70:
30 [v/v]) at a flow rate of 0.8 mL/min. Analytes were
detected at 361 nm.
VB12 was qualitative by HPLC connected to an elec-
trospray mass spectrometer (Micromass Esquire 3000 plus;
Bruker Daltonics, Germany) [15]. The electron spray ion-
ization–mass spectrum (ESI–MS) was performed in posi-
tive mode using 0.01 % (v/v) formic acid in 15 % (v/v)
acetonitrile as mobile phase. Separation was performed at
room temperature at 1 mL/min and the injection volume
was 50 lL. Nitrogen was used as both desolvation gas at a
flow rate of 350 L/h and cone gas at 50 L/h. The desol-
vation temperature was set at 350 °C. The ionization
source was working at 105 °C. Capillary and cone voltages
were 4000 and 180 V, respectively [15, 16].
Identification of the isolates by 16S rDNA
sequencing
Genomic DNA was isolated from the bacterial isolate and
was used as template for polymerase chain reaction (PCR).
Primers used for the amplification of part of 16S rDNA
were 16SF (50 -AGAGTTTGATCCTGGCTCAG-30 ) and
16SR (50 -TACGG TTACCTTGTTACGACTT-30 ). PCR
was carried out in a 50 lL reaction volume containing
50 ng of genomic DNA, 20 pmol of each primer, 1.25 units
of Q5 High-Fidelity DNA polymerase (New England
Biolabs, USA), 200 lM of each dNTPs and 1X PCR buf-
fer. PCR was carried out for 30 cycles in a MycyclerTM
(Bio-Rad, USA) with the initial denaturation at 98 °C for
5 min, cyclic denaturation at 98 °C for 30 s, annealing at
55 °C for 30 s and extension at 72 °C for 1 min with a final
extension of 10 min at 72 °C. The PCR product was
checked by agarose gel electrophoresis and was further
subjected to sequencing. The sequence data was checked
by BLAST analysis [17]. The phylogenetic analysis of the
16S rDNA sequence of the isolate obtained in this study
was conducted with MEGA 5 using neighbor joining
method with 1000 bootstrap replicates [18].
Genome analysis of S. meliloti 320 strain
The genomic DNA of strain S. meliloti 320 was extracted
from purified bacterial isolates using OMEGA Bacterial
DNA kits (Omega bio-tek, Feyou Biotechnology, China).
Sinorhizobium meliloti 320 was completely sequenced
using the sequencing platform on NGS-QC-Toolkit,
deleting overlapping sequences. The cut-off-level (cut off
Qual Score) was set to 30 [19]. The resulting genomic
sequences were determined by BLASTX(V2.2.17) [20].
1529
Searches against protein data bases with the priority order
of non-redundant protein sequences in NCBI(NR) were
carried out by Kyoto Encyclopedia of Genes and Genomes
database (KEGG, Ver. 1.67x, http://www.genome.jp/kegg/
) [21]. Furthermore, searches for Clusters of Orthologous
Groups (COG, http://clovr.org/docs/clusters-of-ortholo
gous-groups-cogs/) [22] were conducted with acritical
E value B1e-3. Based on the results of the protein database
annotation, Blast2GO (https://www.blast2go.com/) [23]
was applied to obtain the functional classification of the
genomic sequences based on gene ontology (GO) terms.
The Nr database was used to perform the GO functional
classification for all genomic sequences.
Optimization of fermentation temperature
The optimal fermentation temperature of strain S. meliloti 320
was tested using the following medium: Sucrose, 60 g L-1;
CSL, 20 g L-1; (NH4)2SO4, 2g L-1; CoCl2, 0.006 g L-1;
KH2PO4, 2.6 g L-1; betaine, 6 g L-1; DMBI, 1 g L-1. Fer-
mentation was carried out in 250 mL-Erlenmeyer flasks
containing 30 mL of medium at an agitation rate of 200 rpm
for 7 days. After 7 days fermentation at 25, 30, 35, 40 and
45 °C, respectively, fermentation broth was used to assay the
concentration of VB12.
Selection of the optimal carbon source and nitrogen
source
Applied media for the determination of the optimal carbon
source contained CSL, 20 g L-1; (NH4)2SO4, 2 g L-1;
CoCl2, 0.006 g L-1; KH2PO4, 2.6 g L-1; betaine, 6 g L-1;
DMBI 1 and 60 g L-1 of one of five carbon sources
(glucose, maltose, sucrose, fructose or starch). VB12-con-
centrations in 7 days fermentation broth were determined
by HPLC as described above.
To determine the most suitable organic nitrogen source,
the following medium was used: Sucrose, 60 g L-1;
(NH4)2SO4, 2 g L-1; CoCl2, 0.006 g L-1; KH2PO4,
2.6 g L-1; betaine, 6 g L-1 and DMBI, 1 g L-1. Organic
nitrogen sources and amounts tested were as following: (1)
2 % yeast powder (YP), 1 % peptone; (2) 2 % yeast extract
(YE), 1 % peptone; (3) 2 % yeast extract powder (YEP),
1 % peptone; (4) 2 % yeast extract concentrate (YEC), 1 %
peptone; (5) 2 % autolyzed yeast powder (AYP), 1 %
peptone; (6) 2 % CSL; (7) 2 % CSL, 1 % peptone.
Experiments were conducted in 250 mL-Erlenmeyer-flasks
containing 30 mL of each medium at 30 °C at an agitation
rate of 200 rpm in rotary shakers for 7 days.
Response surface methodology (RSM)-central compos-
ite design (CCD) was employed to find out the optimum
concentrations of the optimal carbon source and nitrogen
source and to study their interactions [24]. Experimental
123
1530 Bioprocess Biosyst Eng (2016) 39:1527–1537

variable was analyzed at five different levels. An experi-

mental design combining four replicates at the central point
and four axial points (a = 2), with a total of 12 experi-
ments, was used (Table 2). The matrix for this design and
the experimental response values (VB12 concentration) are
shown in Table 2.

Results and discussion

Fig. 2 a The microscope image of S. meliloti 320 dyed by 1 % base
fuchsin; b 6 d fermentation broth of S. meliloti 320 and negative
control
design, model calculation, graph drawing and other data
analysis were all performed on Design Expert software
(Version 8.0.6, Stat-Ease Inc., Minneapolis, USA). Each
Strain selection and identification of VB12
50 bacterial strains were isolated from soil samples and
cultivated in minimum media. Among these strains, only
the color of S. meliloti 320 fermentation broth was red
(Fig. 2b) and the microscope image of S. meliloti 320

Fig. 3 Identification of VB12

compound by LC–MS. A HPLC
chromatogram: fermentation
samples (a, b), negative control
(c) and positive control (d).
B This figure shows the MS
spectrum (e) of purified VB12
(big peak at 12 min shown in a,
b, and d) in relative abundance
against m/z. The MS spectra
allow detection of CN-VB12 at
1355.55 m/z and of
characteristic pseudomolecular
fragments of 147.0859,
359.0980 and 678.2894 m/z,
respectively

123
Bioprocess Biosyst Eng (2016) 39:1527–1537 1531

Fig. 4 Phylogenetic tree of S.

meliloti 320 based on the 16S-
rDNA-sequence. Given
numbers represent confidence
levels of relation showing
Ensifer adhaerens M14 and
Sinorhizobium s. ZJB1101 as
next relatives of S. meliloti 320

Table 1 General information this product, HPLC–ESI–MS/MS was used to analyze the
Features Value
about sequenced S. meliloti 320 fermentation broth.
genome Genome size (bp) 7256751 Electrospray ionization of the cobalamines in a mass
GC-content 63.2 % spectrometer ion source leads to the formation of not only
Plasmid 0 of some molecular adducts, but also a pattern of frag-
Replicons 1 mentation ions (due to a low chemical stability of the
Total genes 6903 cobalamines). Proper selection of a sufficiently abundant
tRNA genes 3 fragment ion helps achieve the required quantification/de-
rRNA genes 5 tection limits. Results of the HPLC–ESI–MS/MS analysis
CDs 6895 from the cell cultures within the MS are shown in Fig. 3B.
Ionization of CN-Cbl using an ESI source leads to the
formation of the ions 678.3 [M?2H]2?, which was consist
with that reported by Szterk [25]. In this study, only CN-
shows it is a rod-shaped strain and the cell diameter is Cbl could be detected as different kinds of vitamin B12
about 1.0–1.2 lm (Fig. 2a). were all translated to CN-Cbl. Lou et al. [16] also reported
The 56 h fermentation broth was used to assay the the formation of multi-charged molecules during B12
amounts of VB12 by HPLC. In compliance with VB12 vitamin ionization. However, some of such adducts were
standards, mass spectroscopic measurements of corre- not found in this study. This is caused by the differences in
sponding HPLC retention peaks revealed spectra covering the electrospray source operation parameters and differ-
VB12-standard spectra (Fig. 3A). This shows that S. meli- ences in the chemical composition of the LC mobile phase.
loti 320 could produce VB12 when grown under aerobic The ionization process might be also affected by the
conditions. To further identification and characterization of composition of the analyzed matrix.

123
1532 Bioprocess Biosyst Eng (2016) 39:1527–1537

Fig. 5 The network of porphyrin and chlorophyll metabolism has shown on KEGG which containing VB12 biosynthetic pathway. Enzyme
classes matching with KEGG analyses are highlighted

Phylogenetic classification of S. meliloti 320 variable regions. This gene is therefore suitable for phylo-
genetic and taxonomic classifications at various levels,
The 16s rDNA gene is a universal characteristic in including intrageneric differentiation [26]. A 1476 bp target
prokaryotes and has both conserved and sufficiently fragment was amplified by PCR using 16S rDNA primers.

123
Bioprocess Biosyst Eng (2016) 39:1527–1537
Fig. 6 a GO functional
classification of genes; b COG
functional classification of
genes
We conducted sequence similarity analysis of 16S rDNA
genes, and compared them with those of reference organ-
isms obtained from GenBank data libraries. The phyloge-
netic analysis by MEGA5 [27] revealed that the isolated S.
meliloti 320 belongs to the species of genus Sinorhizobium/
Ensifer (Fig. 4). Sinorhizobium sp. ZJB1101 and Ensifer
adhaerens M14 represent the most related organisms with
16S rDNA-sequence similarities of 99 %.
1533
Genomic features of S. meliloti 320
Sinorhizobium meliloti 320 hosts one circular chromosome
of 7,256,751 bp with an average GC-content of 63.2 %,
although no further plasmid was present. The genome
comprises 6903 genes and 6895 putative genomic coding
sequences (CDs). The genome includes 3 tRNA genes and
5 rRNA genes in total (Table 1).
123
1534 Bioprocess Biosyst Eng (2016) 39:1527–1537

Figure 5 sums up the network of porphyrin and chloro-

phyll metabolism found via KEGG analysis, including VB12
synthetic pathway. Figure 5 emphasizes the separation of
aerobic and anaerobic pathway both ending in cobalt-(II)-
yrinate-a,c-diamide as precursor of VB12. This figure high-
lights the enzymes and enzymatic clusters (colored green)
involved in ALA, following and further metabolic processes
connected to VB12 pathway by showing the metabolic steps
with regard to genomic analyses. The direct comparison of
the anaerobic and aerobic pathway shows that the presences
of enzymes in the aerobic pathway proving a complete VB12
synthetic pathway present in S. meliloti 320. Simultaneously,
this metabolic map generally announces the importance of
cobalt and iron cations as central components in various
porphyrin and chlorophyll metabolism steps.
The GO analysis (Fig. 6a) shows significantly enriched
functional groups related to catalytic and binding functions,
organization of cellular processes, establishment of local-
ization, and metabolic processes. The COG-analysis of all
genes (Fig. 6b) reveals that the top five categories describe
features of general function (R), amino acid transport and
metabolism (E), carbohydrate transport and metabolism
(G), inorganic ion transport and metabolism (P) and tran-
scription (K). These activities are required for amino acid,
carbohydrate, and energy supply, which in turn leads to
active cell division and related molecular events to facili-
tate VB12 production.
The understanding of the genome information could
provide the opportunities for further studies addressing the
bacterial metabolism and its biotechnological potential.
Biochemical mapping may uncover interfaces, options of
metabolic regulation, and chances to optimize VB12 pro-
duction. Moreover, mentioned metabolic features are also
based on transcriptional regulation. Bio-fermentative VB12
production, interfering porphyrin and chlorophyll metabo-
lism, extensively includes parameters and possibilities
which we have to deal with when declaring efficient VB12
production as object of interest.
Deeper insights into metabolic preferences and behavior
of S. meliloti 320 will provide promising opportunities of
genetic and metabolic engineering enhancing its industrial
application in VB12 production. Vast genetic diversity, and
thus, a variety of different physiological and metabolic
preferences, gives a reason to expand systematic screening
efforts on S. meliloti strains maybe exposing strains of
great impact on industrial and biotechnological application.
Fig. 7 a Effect of temperature on vitamin B12 production by S.
meliloti 320. b Effects of different carbon sources on vitamin B12
Optimization of fermentation temperature production by S. meliloti 320. 1 Glucose, 2 fructose, 3 maltose, 4
sucrose and 5 starch. c Effect of different nitrogen sources on VB12
To optimize fermentation temperature for VB12 produc- production. 1 2 % YP, 1 % peptone; 2 2 % YE, 1 % peptone; 3 2 %
YEP, 1 % peptone; 4 2 % YEC, 1 % peptone; 5 2 % AYP, 1 %
tion, fermentations were conducted at different temperature
peptone; 6 2 % CSL; 7 2 % CSL, 1 % peptone
including 25, 30, 35, 40 and 45 °C. As shown in Fig. 7a,

123
Bioprocess Biosyst Eng (2016) 39:1527–1537 1535

the highest concentration of VB12, 100 ± 2 mg L-1, was
obtained at 30 °C. This temperature was then used in
subsequent experiments.
Selection of carbon source and nitrogen source
To select the optimum carbon source, five different carbon
sources were investigated initially, at pH 7.0, 30 °C and
200 rpm. When using sucrose, 98 ± 2.8 mg L-1 VB12 was
produced, which is the highest one (Fig. 7b). However,
when employing starch as carbon source, VB12 concen-
tration is only 30 mg L-1. Thus, sucrose was then used in
subsequent experiments.
Replacing YP, YE, YEP, YEC or AYP in fermentation
medium with 20.0 g L-1 CSL resulted in VB12 concen-
tration increased from 10 ± 0.8 to 130 ± 3.2 mg L-1
(Fig. 7c). CSL is a cheap organic nitrogen source and is
available in large scale fermentation. Thus, in subsequent
experiments CSL was chosen as suitable nitrogen source
and its concentration was further to be optimized.
Optimization of the concentration of sucrose
and CSL using RSM
The concentrations of sucrose and CSL were optimized to
find the best combination of them for maximum VB12
production. 1 % peptone was added to each experiment to
increase VB12 production. The experimental data in
Table 2 were correlated as a second-order polynomial
model by nonlinear regression. The empirical relationship
between variables and responses was expressed by the
following second-order polynomial equation:
Y ¼ 126:46 þ 9:58A þ 18:08Bþ8:25AB  11:19A2
 9:94B2
where Y is the predicted response of VB12 production (mg
L-1), and A and B are the coded values of sucrose and
CSL, respectively.
The sufficiency of the second-order model was checked
by the analysis of variance (ANOVA) presented in Table 3.
The F value of 5.37 implies that the model is significant, as
there is only a 3.21 % probability that a large F value could
occur due to noise. The model presents a relatively high
determination coefficient (R2 = 0.8172), explaining
81.7 % of the variability in the response. The P values
were used to identify the effects of each factor on VB12
production. A P value below 0.05 indicates that the model
terms are significant. In this case, A2 (sucrose 9 sucrose),
B (CSL) and B2 (CSL 9 CSL) are significant (Table 4).
The interaction between these two factors is not significant.
The location of the optimum was determined to be
A = -0.901, and B = 1.283, obtained by the differentia-
tion of the quadratic model given by equation above. The
corresponding uncoded values were: 69.0 g L-1 sucrose

Table 2 Experimental design

Run A (coded value) B (coded value) A (sucrose, g L-1) B (CSL, g L-1) VB12 (mg L-1)
and results of RSM
1 -1 -1 50 15 67
2 1 -1 70 15 78
3 -1 1 50 25 83
4 1 1 70 25 127
5 -2 0 40 20 75
6 2 0 80 20 105
7 0 -2 60 10 57
8 0 2 60 30 133
9 0 0 60 20 130
10 0 0 60 20 145
11 0 0 60 20 126
12 0 0 60 20 138

Table 3 ANOVA of the

Source Sum of squares Degrees of freedom Mean square F value P value
quadratic model
Model 8893.54 5 1778.71 5.37 0.0321
Residual 1989.12 6 331.52
Lack of fit 1774.37 3 591.46 8.26 0.0582
Pure error 214.75 3 71.58
Cor total 10882.67 11

123
1536 Bioprocess Biosyst Eng (2016) 39:1527–1537

Table 4 Regression coefficients and significances obtained by RSM
Factor Coefficient estimate F value P value
Intercept 126.46
A-sucrose 9.58 3.32 0.1181
B-CSL 18.08 11.84 0.0138
AB 8.25 0.82 0.3998
A2 -11.19 8.05 0.0296
B2 -9.94 6.35 0.0452
and 26.4 g L-1 CSL. The predicted optimal VB12 pro-
duction according to these values was 142.4 mg L-1.
To confirm the adequacy of the model to predict max-
imum VB12 production, a verification experiment using the
optimum medium composition was performed. The three
replicate experiments yielded an average VB12 concentra-
tion of 140 ± 4.2 mg L-1. The good agreement between
the predicted and experimental results verified the validity
of the model and the existence of an optimal point.
Conclusion
A new unicellular vitamin B12-producing strain, S. meliloti
320, was isolated and identified from soil samples.
Although there are several reports about VB12 production
by various aerobic bacteria, this strain is a competitive
candidate for VB12 production due to its high VB12 pro-
duction. VB12 production reached 140 mg L-1 after 7 days
fermentations at 30 °C using 69.0 g L-1 sucrose and 26.4 g
L-1 CSL as carbon and nitrogen sources, respectively.
Deeper insights into genomic information and behavior of
S. meliloti 320 will provide promising opportunities of
genetic and metabolic engineering enhancing its industrial
application in VB12 production.
Acknowledgments We are grateful for the support from State Key
Development Program for Basic Research of China (973 Program,
2013CB733600), National Nature Science Foundation of China
(31370089) and Key Laboratory of Systems Microbial Biotechnol-
ogy, Tianjin Institute of Industrial Biotechnology, Chinese Academy
of Sciences.
References
1. Campbell GR, Taga ME, Mistry K, Lloret J, Anderson PJ, Roth
JR, Walker GC (2006) Sinorhizobium meliloti bluB is necessary
for production of 5,6-dimethylbenzimidazole, the lower ligand of
B12. Proc Natl Acad Sci USA 103(12):4634–4639
2. Zhu X, Wang XF, Zhang C, Wang XQ, Gu Q (2015) A ribos-
witch sensor to determine vitamin B12 in fermented foods. Food
Chem 175:523–528
3. Sally P, Stabler M (2014) Vitamin B12 deficiency. New Engl J
Med 368:149–160
4. Xia W, Chen W, Peng WF, Li KT (2015) Industrial vitamin B
production by Pseudomonas denitrificans using maltose syrup
and corn steep liquor as the cost-effective fermentation sub-
strates. Bioprocess Biosyst Eng 38(6):1065–1073
5. Martens JH, Barg H, Warren MJ, Jahn D (2002) Microbial pro-
duction of vitamin B12. Appl Microbiol Biotechnol
58(3):275–285
6. Kang Z, Zhang JL, Zhou JW, Qi QS, Du GC, Chen J (2012)
Recent advances in microbial production of delta-aminolevulinic
acid and vitamin B12. Biotechnol Adv 30(6):1533–1542
7. Deptula P, Kylli P, Chamlagain B, Holm L, Kostiainen R,
Piironen V, Savijoki K, Varmanen P (2015) BluB/CobT2 fusion
enzyme activity reveals mechanisms responsible for production
of active form of vitamin B12 by Propionibacterium freudenre-
ichii. Microb Cell Fact 14(1):186
8. Anderson PJ, Lango J, Carkeet C, Britten A, Krautler B, Ham-
mock BD, Roth JR (2008) One pathway can incorporate either
adenine or dimethylbenzimidazole as an a-axial ligand of B12
cofactors in Salmonella enterica. J Bacteriol 190(4):1160–1171
9. Strohm TO, Griffin B, Zumft WG, Schink B (2007) Growth
yields in bacterial denitrification and nitrate ammonification.
Appl Environ Microbiol 73(5):1420–1424
10. Ko Y, Ashok S, Ainala SK, Sankaranarayanan M, Chun AY, Jung
GY, Park S (2014) Coenzyme B12 can be produced by engineered
Escherichia coli under both anaerobic and aerobic conditions.
Biotechnol J 9(12):1526–1535. doi:10.1002/biot.201400221
(Epub 2014 Oct 10)
11. Gibson KE, Kobayashi H, Walker GC (2008) Molecular deter-
minants of a symbiotic chronic infection. Ann Rev Genet
42:413–441
12. Kliewer M, Evans HJ (1962) Effect of cobalt deficiency on the
B12 coenzyme content of Rhizobium meliloti. Arch Biochem
Biophys 97:427–429
13. Wang P, Wang YS, Liu YD, Shi H, Su ZG (2012) Novel in situ
product removal technique for simultaneous production of pro-
pionic acid and vitamin B12 by expanded bed adsorption biore-
actor. Bioresour Technol 104:652–659
14. Cheng X, Chen W, Peng WF, Li KT (2014) Improved vitamin
B12 fermentation process by adding rotenone to regulate the
metabolism of Pseudomonas denitrificans. Appl Biochem
Biotechnol 173(3):673–681
15. Madhu AN, Giribhattanavar P, Narayan MS, Prapulla SG (2010)
Probiotic lactic acid bacterium from kanjika as a potential source
of vitamin B12: evidence from LC–MS, immunological and
microbiological techniques. Biotechnol Lett 32(4):503–506
16. Luo XB, Chen B, Ding L, Tang F, Yao SZ (2006) HPLC-ESI-MS
analysis of vitamin B12 in food products and in multivitamins-
multimineral tablets. Anal Chim Acta 562(2):185–189
17. Zhang Z, Schwartz S, Wagner L, Miller W (2000) A greedy
algorithm for aligning DNA sequences. J Comput Biol
7(1–2):203–214
18. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S
(2011) MEGA5: molecular evolutionary genetics analysis using
maximum likelihood, evolutionary distance, and maximum par-
simony methods. Mol Biol Evol 28(10):2731–2739
19. Li R, Zhu H, Ruan J, Qian W, Fang X, Shi Z, Li Y, Li S, Shan G,
Kristiansen K, Li S, Yang H, Wang J, Wang J (2010) De novo
assembly of human genomes with massively parallel short read
sequencing. Genome Res 20(2):265–272
20. Altschul S, Madden T, Schäffer A, Zhang J, Zhang Z, Miller W,
Lipman D (1997) Gapped BLAST and PSI-BLAST: a new gen-
eration of protein database search programs. Nucleic Acids Res
25(17):3389–3402
21. Kanehisa M, Araki M, Goto S, Hattori M, Hirakawa M, Itoh M,
Katayama T, Kawashima S, Okuda S, Tokimatsu T, Yamanishi Y

123
Bioprocess Biosyst Eng (2016) 39:1527–1537
(2008) KEGG for linking genomes to life and the environment.
Nucleic Acids Res 36(Database issue):480–484
22. Zhang XW, Ye ZW, Wang TK, Xiong HR, Yuan XL, Zhang ZG,
Yuan YL, Liu Z (2014) Characterization of the global tran-
scriptome for cotton (Gossypium hirsutum L.) anther and devel-
opment of SSR marker. Gene 551(2):206–213
23. Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles
M (2005) Blast2GO: a universal tool for annotation, visualization
and analysis in functional genomics research. Bioinformatics
21(18):3674–3676
24. Dong HN, Zhao XM, Ma YY, Zhang MH (2012) Optimization of
a synthetic medium for ethanol production by xylose-fermenting
Zymomonas mobilis using response surface methodology. Chin
Sci Bull 57(28–29):3782–3789
1537
25. Szterk A, Roszko M, Malek K, Czerwonka M, Waszkiewicz-
Robak B (2012) Application of the SPE reversed phase HPLC/
MS technique to determine vitamin B12 bio-active forms in beef.
Meat Sci 91(4):408–413
26. Seemüller E, Schneider B, Mäurer R, Ahrens U, Daire X, Kison
H, Lorenz K, Firrao G, Avinent L, Sears B, Stackebrandt E
(1994) Phylogenetic classification of phytopathogenic mollicutes
by sequence analysis of 16S ribosomal DNA. Int J Syst Bacteriol
44(3):440–446
27. Ji F, Liu Y, Hao R, Li G, Zhou Y, Dong R (2014) Biomass
production and nutrients removal by a new microalgae strain
Desmodesmus sp. in anaerobic digestion wastewater. Bioresour
Technol 161:200–207
123
Bioprocess & Biosystems Engineering is a copyright of Springer, 2016. All Rights Reserved.