Abstract

   Corynebacterium glutamicum (C. glutamicum) is a favorable host cell for
   the production of recombinant proteins, such as important enzymes and
   pharmaceutical proteins, due to its excellent potential advantages.
   Herein, we sought to systematically explore the influence of
   recombinant protein expression on the transcription and metabolism of
   C. glutamicum. Two C. glutamicum strains, the wild-type strain and an
   engineered strain expressing enhanced green fluorescent protein (EGFP),
   were cultured in parallel in 5-L bioreactors to study the change in
   metabolism in the process of EGFP expression. The results revealed that
   EGFP expression had great effects on the growth and metabolism of C.
   glutamicum and contributed to metabolism-like anaerobic conditions as
   follows: glycolysis was enhanced, the TCA cycle was shunted, and Glu,
   Val, Met, lactate and acetate were accumulated to produce sufficient
   ATP for EGFP production and transfer. Many differentially expressed
   genes related to ribosomal protein, transcriptional regulators, and
   energy metabolism were found to be expressed in the presence of EGFP,
   laying the foundation for identifying genomic loci to change the flow
   of the host cell metabolism to improve the ability of expressing
   foreign proteins in C. glutamicum.

Introduction

   Since 1957, when the Gram-positive bacterium Corynebacterium glutamicum
   (C. glutamicum) was isolated due to its ability to excrete large
   amounts of L-glutamate [[36]1], C. glutamicum has been thoroughly
   investigated and used as a generally regarded-as-safe organism in the
   fermentation industry. Within the last decades, C. glutamicum was
   proven to be an excellent production platform for amino acids and
   organic acids, and a promising alternative bacterial host for the
   expression of recombinant proteins due to its unique characteristics.
   First, C. glutamicum is non-pathogenic and does not produce endotoxins.
   Furthermore C. glutamicum could secrete properly folded proteins into
   the culture medium and lacks detectable extracellular hydrolysis
   activity, activities that increase the subsequent purification
   efficiency [[37]2]. With these advantages, C. glutamicum has been
   widely used to express recombinant proteins such as staphylococcal
   nuclease [[38]3], transglutaminase [[39]4], human epidermal growth
   factor [[40]5], antibody fragment [[41]6], green fluorescent protein
   (GFP) [[42]7] and α-amylase [[43]8], et al.

   When C. glutamicum expresses recombinant proteins, it has several
   intrinsic disadvantages, such as a low transformation efficiency, few
   available expression vectors [[44]6], and a lower yield than that of
   Escherichia coli. Tremendous efforts have been put forth into the
   genetic modification and cultivation of C. glutamicum to avoid these
   issues—for example, the development of sophisticated tools for genetic
   manipulation, the engineering of natural and endogenous promoters and
   signal peptides, the optimization of codon usage, and the knockout of
   some proteases to reduce the degradation of recombinant proteins
   [[45]9–[46]13]. In addition, recombinant protein expression is also an
   ATP-consuming process, and an adequate supply of ATP could
   appropriately improve the expression level of recombinant proteins
   [[47]14]. Due to the complex process of heterologous protein expression
   and energy metabolism of C. glutamicum, it is not easy to improve the
   production of heterologous protein simply through increasing the ATP
   level but through identifying the positively correlated genes between
   them.

   In this study, to systematically explore the effect of enhanced green
   fluorescent protein (EGFP) expression on the genetic regulation and
   metabolism of C. glutamicum, we compared C. glutamicum Egfp and C.
   glutamicum BZH 001 in a 5-L bioreactor through RNA-seq. Subsequently,
   the mechanistic correlation among these changes and EGFP expression was
   studied to understand the metabolism mechanism in the process of
   heterologous protein expression and to identify key nodes to reform the
   genomic loci to improve the ability to express foreign proteins. In
   general, the study of metabolism and transcriptome regulation in the
   heterologous protein expression of C. glutamicum has high academic
   value and broad application prospects.

Materials and methods

Strains and cultivation

   The wild-type strain used in this work was C. glutamicum CGMCC1.15647
   (C. glutamicum BZH 001) donated by Zhangjiagang Huachang Pharmaceutical
   Co., Ltd. (Zhangjiagang, China). Its derivatives C. glutamicum
   pXMJ19-EGFP (C. glutamicum EGFP) was constructed as the following. EGFP
   gene was amplified by PCR using F primer
   (5’-CCAAGCTTAAAGGAGGACAACTAATGGTGAGCAAGGGCG-3’) and R primer
   (5’-CCCGGATCCTTACTTGTACAGCTCGTCCATG-3’) with plasmid PEGFP-N1 as the
   template. The amplified DNA fragment was introduced into the PXMJ19 by
   Hind III and BamH I digestion and ligation. The lacIq gene was removed
   from PXMJ19 to make it a constitutive expression vector. The resulting
   plasmid PXMJ19-EGFP was introduced into C. glutamicum BZH 001. The EGFP
   gene was constitutive express by tac promoter witha conserved SD
   sequence (aaaggagga). The physical map of PXMJ19-EGFP was shown in
   [48]S1 Fig [[49]15].

   The strain was grown on seed medium (25 g of glucose, 30 g of corn
   syrup, 20 g of (NH[4])[2]SO[4], 1 g of MgSO[4], and 1 g of KH[2]PO[4]
   per liter of distilled H[2]O; pH 6.8) in a shaker for 12 h at 30°C and
   220 rpm. Next, the seed was transferred to the second seed medium in a
   baffled-bottom flask at 30°C, pH 6.8, for 48 h before transferring to
   fermentation medium in a 5-L fermenter (Applikon EZ-control,
   Netherlands). Fed batches were conducted in parallel with three
   biological replicates of C. glutamicum BZH 001, and C. glutamicum EGFP
   under 30% dissolved oxygen (DO). Glucose (300 g/L) was added into the
   fermenter at the rate of 10 mL/h 16 h after inoculation to avoid
   nutrient shortage. DO calibration and control: In this experiment, 100%
   DO was calibrated after equilibration for 5 h with an aeration rate of
   3 L/min and an agitator speed of 400 rpm at 30°C. The DO level
   (percentage of air saturation) was controlled at constant value by
   varying the agitation from 400 to 1000 rpm and supplying pure oxygen
   through a solenoid valve. The DO value was allowed to fluctuate by less
   than 4%. The medium was composed of 30 g of glucose, 15 g of corn
   syrup, 20 g of (NH[4])[2]SO[4], 1 g of MgSO[4], and 1 g of KH[2]PO[4]
   per liter of distilled H[2]O at pH 6.8. The samples collected at 20 h
   from three independent batches were used for RNA sequencing and
   quantitative real-time RT-PCR analysis. Samples for metabolism analysis
   were collected every 4 h during the entire fermentation process. All
   samples were quenched in liquid nitrogen immediately after collection
   and were then stored at -80°C before analysis.

Biomass and concentration of glucose

   Biomass formation was followed by measuring the OD[600] and by
   determining the cell dry weight. The concentration of glucose was
   measured using the 3, 5-dinitrosalicylic acid (DNS) method [[50]16].

HPLC analyses

   The determination of amino acids and organic acids was performed as
   described previously [[51]17, [52]18] by 1260 Agilent system (Agilent
   Technologies Inc., USA) and SHIMADZU HPLC system (SHIMADZU Corp.,
   Japan).

Quantification of enzyme activities

   Quantification of the pyruvate carboxylase (PC), isocitrate
   dehydrogenase (ICDH), glutamate dehydrogenase (GDH) activities was
   performed according to modified protocols. The protein concentration
   was quantified using the Bradford assay. Specific activities were
   calculated as follows: a[s] = a/TP [[53]17].

Quantification of energy

   ATP was measured using an ATP Assay Kit (Beyotime, China).

Transcriptome analysis by RNA-seq

   The harvesting of total RNA and the RNA-Seq protocol were described
   previously [[54]19]. The clean reads were exported in the FASTQ format
   and were deposited in the National Center for Biotechnology Information
   (NCBI) Sequence Read Archive (SRA) with the accession numbers
   [55]GSE77502/[56]GSE87077. The clean reads were aligned to the
   reference sequences (C. glutamicum ATCC13032 genome) by
   SOAPaligner/SOAP2 [[57]20].

   The alignment data were used to calculate the distribution of reads on
   reference genes and to perform coverage analysis. After the alignment
   results passed QC, downstream analysis was performed, including gene
   expression level analysis, differential expression analysis, pathway
   enrichment analysis, and gene ontology (GO) enrichment analysis. RPKM
   (reads per kb per million reads) was used for the gene expression level
   calculation [[58]19, [59]21, [60]22].

Quantitative real-time PCR

   RNA was quantified using a NanoDrop ND-500 spectrophotometer and was
   used for cDNA synthesis. Real-time RT-PCR was performed using an ABI
   stepone Plus Real-Time PCR System and the SYBR RT-PCR Kit. The 16s RNA
   gene was used as an internal control. Eight genes were chosen for
   expression analysis, and the primers are shown in [61]S1 Table. The
   amplification conditions were as follows: 30 s at 95°C, 40 cycles of 5
   s at 95°C, 30 s at 60°C, and 20 s at 72°C. The relative gene expression
   levels were calculated using the 2^-ΔΔCt method. Each experiment was
   repeated three times.

Method of MVDA (PCA, OPLS-DA), Heatmap

   MVDA was carried out by SIMCA-P 14(MKS UmetricsAB, Sweden) as the
   method described previously [[62]19, [63]23].

   Heatmap was performed by cluster software [[64]24] and Java Treeview
   software [[65]25]. The RPKM was deal by hierarchical clustering and
   Euclidean distance.

Results

Transcriptome sequencing (RNA-seq) output and expression annotation

   In this study, three fermentation batches were conducted using C.
   glutamicum BZH 001 and C. glutamicum EGFP, respectively. Transcriptome
   analysis was performed on C. glutamicum BZH 001 and C. glutamicum EGFP
   to elucidate the relationship between heterologous protein expression
   and metabolism. Samples harvested at 20 h in fermenters were used for
   RNA-seq. At that point, the cells remained at the early stage of the
   stable phase. We studied the changes in gene expression from the
   samples taken after 20 h of fermentation from cultures under 30% DO
   concentrations. Each sample had three independent reproducible
   biological replicates; the average coefficient of variation (CV) was
   lower than 2.50% ([66]S2 Table). After the QC step (analysis of base
   composition and quality), the raw reads were qualified and filtered
   into clean reads and then were mapped to the C. glutamicum ATCC13032
   genome (The accession number is [67]NC_003450.3). The total reads
   generated in each sample library ranged from 19,321,622 to 27,005,014.
   In total, approximately more than 16,300,000 reads were uniquely mapped
   to the genome, and their percentage in total reads was above 84%
   ([68]S3 Table). Next, NOIseq was used to deal with clean data of the
   three duplicates to identify the differentially expressed genes (DEGs).
   An absolute value of the log2 ratio > 1 and FDR<0.05 were used to
   select genes with significant differential expression. When the C.
   glutamicum BZH 001 group was set as the control, 244 genes were
   significantly down-regulated, and 319 genes were up-regulated; 14 genes
   with |log2 ratio| > 4 were analyzed here. Four of the 14 DEGs
   (NCgl1413, NCgl2845, NCgl2837 and NCgl0910) were related to a
   hypothetical protein; thus, the focus was on the others. Among the 10
   DEGs, NCgl0055 had a potential function in transcription regulation,
   and 5 genes had no gene annotation; thus, the other 4 DEGs (NCgl0487,
   NCgl0909, NCgl0833 and NCgl0303) were further analyzed. NCgl0487,
   encoding the 50S ribosomal protein L3 (rplC), was almost 95-fold
   down-regulated. rplC is a ribosomal protein involved in the translation
   elongation process, and its down-regulation indicates that fewer
   ribosomes are being produced to translate EGFP protein. NCgl0833
   encoding the 50S ribosomal protein L33 (rpmG) was also down-regulated
   by 24.18-fold, and it was identified as a zinc-binding domain. NCgl0909
   encoding the ABC transporter ATPase was significantly up-regulated. ABC
   transporters are members of a transport system superfamily, and ATPase
   is a subunit of ABC transporters that utilizes the energy of ATP
   binding and hydrolysis to transport various substrates across cellular
   membranes. The up-regulation of this gene in the C. glutamicum EGFP
   group indicated that the transmembrane transportation of some
   substrates was activated, possibly to transport EGFP. NCgl0303 encoding
   a cold shock protein was up-regulated. Shock protein, part of the CspA
   family, is the major cell surface protein, along with CspB. It was
   investigated that CspB mutation in C. glutamicum may increase the
   secretion of the antibody Fab fragment [[69]13]; thus, up-regulated
   CspA may perform similarly in the C. glutamicum EGFP group.

Analysis of Gene Ontology (GO) functional enrichment and pathway enrichment

   GO enrichment was performed to identify the potential function of the
   DEGs of the C. glutamicum EGFP group. The DEGs were annotated using GO
   annotations of GO term findings
   ([70]https://david.ncifcrf.gov/gene2gene.jsp). Annotated genes were
   enriched with a corrected p-value ≤0.05. The DEGs could be categorized
   into two functional groups belonging to two main GO ontologies:
   biological processes (4) and molecular functions (2). The GO
   annotations belonging to the biological processes were
   GO:0009401-phosphoenolpyruvate-dependent sugar phosphotransferase
   system, GO:0006355-regulation of transcription, DNA-templated,
   GO:0006096-glycolytic process, and GO:0006351-transcription,
   DNA-templated. The molecular functions of GO ontologies were
   GO:0008199-ferric iron binding and GO:0035731-dinitrosyl-iron complex
   binding ([71]Fig 1A). There were 32 genes clustered in transcription
   and 22 genes in the regulation of transcription. Each DEG was subjected
   to pathway analysis using the KEGG (Kyoto Encyclopedia of Genes and
   Genomes) database ([72]http://www.kegg.jp/kegg/pathway.html) to explore
   the biological implications. In addition, we generated a scatter plot
   of the KEGG enrichment results ([73]Fig 1B). A smaller rich factor
   indicates greater intensiveness, and a larger plot indicates more
   enrichment of genes. The p-value ranged from 0 to 0.056. [74]Fig 1
   shows the enrichment pathways of the C. glutamicum EGFP group.
   Specifically, the pathway terms enriched with the largest number of
   DEGs annotated were microbial metabolism in diverse environments; it
   was an important pathway of energy metabolism, substance metabolism and
   transcription and there were 46 genes annotated in this pathway.
   Heterologous protein expression needs many metabolic precursors and/or
   cofactors; thus, the pathway terms of Glycolysis/Gluconeogenesis,
   Carbon metabolism, Pentose phosphate pathway and Pyruvate metabolism
   were enriched. These pathways are important pathways to synthesize all
   required amino acid and ATP necessary to form substrates of the target
   protein. Five genes of the Phosphotransferase system (PTS) were
   enriched, and the p-value of this pathway was 0.0036; PTS could
   efficiently transport different sugars. Anda et al. replaced the
   glucose PTS with alternate glucose transport activity on growth
   kinetics and acetate accumulation to improve the production of
   recombinant proteins [[75]26]. Thus, changing the activity of PTS is an
   important method to improve the cellular performance in recombinant
   protein production systems.

Fig 1. GO classification and pathway enrichment.

   [76]Fig 1
   [77]Open in a new tab

   A: GO classification of different expression genes. B: Pathways
   enrichment of different expression genes.

Expression of the key genes of C. glutamicum BZH 001 and C. glutamicum EGFP

   DEGs were identified at a false discovery rate (FDR) lower than 0.05
   and an absolute value of the log2 ratio≥1 was set as the threshold.
   However, many critical genes regulated by EGFP expression may be missed
   through such a screening method. Thus, we applied MVDA to further
   analyze the RPKM data, and OPLS-DA was performed to identify the
   critical genes. The S-plot model was used to screen the DEGs of the
   groups, and the data analysis method of Par scaling was adopted to
   identify biomarkers. The p1-axis describes the influence of each
   X-variable on the group separation, and the p(corr)1-axis represents
   the reliability of each X-variable to accomplish the group separation.
   Fifty-seven genes were obtained as potential biomarkers from the
   S-plots, and the venn diagram showed that there were 29 common critical
   genes from MVDA and DEG analysis of C. glutamicum EGFP compare with
   those of C. glutamicum BZH 001 ([78]Fig 2, [79]S4 Table). There were 10
   genes that were hypothetical proteins. Eleven genes encoded ribosomal
   proteins (NCgl1304, NCgl0538, NCgl1901, NCgl2261, NCgl0518, NCgl0515,
   NCgl0495, NCgl0487, NCgl0833, NCgl1325, NCgl0488); however, most of
   them were down-regulated under C. glutamicum EGFP, indicating that the
   translation process was suppressed with EGFP expression. There were two
   genes (NCgl0303, NCgl1526) up-regulated in C. glutamicum EGFP. NCgl0303
   encoding cold shock protein had been discussed above. NCgl1526 encoding
   glyceraldehyde-3-phosphate dehydrogenase, a house-keeping gene, was
   up-regulated by 6.61-fold. With the up-regulation of NCgl1526, more ATP
   and NADH were generated. The other genes are related to the ETC
   pathway, ABC transporter system, transcriptional regulator and
   transcription: NCgl2115 encoding cytochrome C oxidase subunit II;
   NCgl2060 encoding ABC transporter ATPase; NCgl2779 encoding esterase;
   NCgl1109 encoding helicase; NCgl2897 encoding starvation-inducible
   DNA-binding protein; NCgl1504 encoding transcriptional regulator. Most
   of them were related to energy metabolism; however, their relationship
   with recombinant protein expression should be further studied.

Fig 2. Heat map and Venn analysis of critical genes from MVDA and DEG.

   [80]Fig 2
   [81]Open in a new tab

   A: Heatmap of 29 genes from MVDA, G1, G2 and G3 mean three replicates
   of C. glutamicum EGFP; W1, W2 and W3 mean three replicates of C.
   glutamicum BZH 001; B: Venn analysis of MVDA and DEG.

Analysis of the growth properties of C. glutamicum BZH 001 and C. glutamicum
EGFP

   The expression of heterologous genes is usually a strong burden to the
   host cells, resulting in a decline in the biomass of bacteria. [82]Fig
   3 shows the comparison of C. glutamicum BZH 001 and C. glutamicum EGFP
   concerning the aspects of biomass, cell dry weight (CDW) and sugar
   consumption (the EGFP expression level was shown in [83]S2 Fig). Both
   strains grew fast in the initial fermentation stage. The biomass of C.
   glutamicum BZH 001 reached its maximum (OD[600nm] = 46.44 at 18 h and
   then declined to 29.70 at 48 h), whereas the biomass of C. glutamicum
   EGFP reached its maximum at 24 h of 43.92 and fell to 31.47 at 48 h. It
   was obvious that the heterologous protein expression inhibited cell
   growth at the early stage because C. glutamicum EGFP grew poorly
   compared with C. glutamicum BZH 001, with a lower max-OD[600] and
   longer time to reach the max biomass ([84]Fig 3A). Therefore, it can be
   reasonably presumed that heterologous protein expression competed with
   the host-cell growth for the precursors of the engineered strain,
   resulting in a lower cell growth rate. [85]Fig 3B shows the CDWs of
   different strains. The CDWs of C. glutamicum EGFP were lower than those
   of C. glutamicum BZH 001 and were kept stable in the middle to late
   stages. The CDWs of both strains arrived at the maximum at 28 h. We
   further analyzed the consumption of sugar and found that the sugar
   consumption rate of C. glutamicum EGFP was lower before 20 h; however,
   after 20 h, it was higher than that of C. glutamicum BZH 001 ([86]Fig
   3C). The results indicated that the expression of foreign proteins not
   only affected the cell growth but also increased glucose consumption to
   supply more energy for protein synthesis. The NCgl1858 encoding
   phosphoenolpyruvate-protein kinase, which is a critical gene in the PTS
   system, was up-regulated by 2.98-fold, and NCgl1305 and NCgl2553
   encoding the phosphotransferase system IIC component were up-regulated
   by 4.54 and 3.86-fold, respectively, both consistently showing the up
   absorption and transfer of glucose. Increased glucose consumption
   during the anaerobic condition generally has been considered to produce
   ATP, and a previous study also reported that increased TCA cycle flux
   and ATP production were shown in heterologous protein expression in
   Pichia pastoris [[87]27].

Fig 3. Biomass, CDW and sugar concentration of C. glutamicum BZH 001 and C.
glutamicum EGFP during 48 h of fermentation.

   [88]Fig 3
   [89]Open in a new tab

Analysis of the concentrations of amino acids and organic acids of C.
glutamicum BZH 001 and C. glutamicum EGFP

   Many studies have investigated the metabolic pathway for the production
   of heterologous proteins and/or biomass. The models to express
   different heterologous proteins were analyzed; these models strongly
   depended on the amino acid composition of the expressed protein, which
   has a profound effect on the number and identity of possible pathways
   for the production of these proteins [[90]27, [91]28]. Vijayasankaran
   et al. found that the amino acid composition of heterologous proteins
   had a profound impact on the predicted elementary flux modes leading to
   high protein production in Escherichia coli [[92]29]. Most of these
   studies were focused on metabolite drain from the central metabolism
   for each required amino acid and the energy requirements for the
   production of each peptide bond. However, we investigated the component
   of different amino acids and organic acids when EGFP was expressed in
   C. glutamicum. [93]Table 1 shows the concentration of amino acids and
   organic acids of two strains at 48 h. Both the intracellular and
   extracellular contents of glutamic acid and alanine were higher than
   those of other amino acids. The extracellular glutamic acid content of
   both strains was higher than the intracellular production, reaching
   2.93 and 3.46 g·L^-1, respectively. Both the intracellular and
   extracellular content of glutamic acid of C. glutamicum EGFP were
   higher than those of C. glutamicum BZH 001. Moreover, the extracellular
   contents of Gly, Val and Met were detectable in C. glutamicum EGFP,
   whereas little was detectable in C. glutamicum BZH 001, which may be
   caused by the up-regulation of glycolysis and TCA shunting with EGFP
   expression. The lactic acid and acetic acid contents were greater in C.
   glutamicum EGFP than in C. glutamicum BZH 001, which is generally
   considered an imbalance between the glucose uptake and up-regulation of
   glycolysis and the limited activity of the TCA cycle [[94]30].
   Moreover, acetate has adverse effects on the synthesis of DNA, RNA,
   proteins, and recombinant protein could be affected by acetate even at
   concentrations as low as 0.5 g/L [[95]28, [96]31]. Six genes involved
   in glycolysis (NCgl1202 encoding 6-phosphofructokinase, NCgl1526
   encoding pyruvate kinase, NCgl2008 encoding glyceraldehyde-3-phosphate
   dehydrogenase and NCgl2673 encoding fructose-bisphosphate aldolase,
   NCgl1525 encoding phosphoglycerate kinase, NCgl2008 encoding pyruvate
   kinase) were up-regulated in C. glutamicum EGFP. However, the gene
   NCgl0355 encoding dihydrolipoamide dehydrogenase was down-regulated in
   C. glutamicum EGFP, indicating that the TCA pathway was down-regulated,
   resulting in an anaerobic-like metabolite. To gain sufficient ATP,
   glycolysis was up-regulated, but the TCA cycle was down-regulated due
   to the lack of an electron acceptor [[97]32]. The changed contents of
   amino acids and organic acids may be related to the amino acid
   composition of EGFP and the demand of ATP in C. glutamicum EGFP.

Table 1. Concentrations of amino acids and organic acids of C. glutamicum BZH
001 and C. glutamicum EGFP at 48 h.

   Name Strains Intracellular content (g·g^-1) Extracellular content
   (g·L^-1)
   Glu BZH001 0.44±0.012 2.93±0.008
   EGFP 1.26±0.036 3.46±0.011
   Ala BZH001 1.08±0.021 1.70±0.017
   EGFP 0.09±0.007 1.35±0.013
   Gly BZH001 — —
   EGFP — 0.31±0.004
   Ile BZH001 — —
   EGFP 0.08±0.003 —
   Val BZH001 — —
   EGFP — 0.76±0.014
   Met BZH001 — —
   EGFP — 1.15±0.021
   Lac BZH001 — 0.06±0.003
   EGFP 0.21±0.018 1.38±0.021
   Ace BZH001 — 0.41±0.17
   EGFP 0.32±0.016 1.76±0.012
   [98]Open in a new tab

   Abbreviations: Glu: glutamic acid; Ala: Alanine; Gly: Glycine; Ile:
   Isoleucine; Val: Valine; Met: Methionine; Lac: Lactate; Ace: Acetate;
   BZH001: C. glutamicum BZH 001; EGFP: C. glutamicum EGFP

Analysis of the enzyme activities of C. glutamicum BZH 001 and C. glutamicum
EGFP

   Enzyme activity is one of the symbols reflecting the intensity of
   metabolic pathways. The regulation of enzymes directly affects the
   regulation of metabolic flux of the path, and it is the most basic
   metabolic regulation. We can indirectly assess the strength of the
   metabolic pathway by detecting enzyme activities. Enzyme activities
   were determined for selected carboxylating and decarboxylating
   reactions of the central metabolism of C. glutamicum BZH 001 and C.
   glutamicum EGFP including the TCA cycle and glutamate synthesis. One of
   the most important anaplerotic reactions was catalyzed by pyruvate
   carboxylase. [99]Fig 4A shows a little higher pyruvate carboxylase (PC)
   activity of C. glutamicum BZH001 than C. glutamicum EGFP at 20 h.
   However, after 20 h, the PC activity of C. glutamicum BZH001 grew more
   quickly than that of C. glutamicum EGFP, reaching 0.1 U·mg^-1. NCgl0659
   encoding PC was 2.17-fold higher than C. glutamicum EGFP from RNA-seq
   data. PC catalyzes the pyruvate acid to oxaloacetic acid reaction, and
   its increased activity supplies more oxaloacetic acid for the TCA cycle
   at the logarithmic growth phase. However, this process could consume
   ATP; thus, it was necessary to down-regulate PC activity with EGFP
   expression. However, the up-regulation of NCgl1523 (encoding
   phosphoenolpyruvate carboxylase) in C. glutamicum EGFP may be another
   way to supply oxaloacetic acid for the TCA cycle and improve the
   proportion of carbon flux through this anaplerotic reaction, reducing
   acetate accumulation and improving protein production [[100]28].

Fig 4. The comparison of specific enzyme activities between C. glutamicum BZH
001 and C. glutamicum EGFP.

   [101]Fig 4
   [102]Open in a new tab

   A: PC; B: ICDH; C: GDH.

   Isocitrate dehydrogenase (ICDH) is also an important enzyme in the TCA
   cycle that catalyzes the generation of α-ketoglutarate from isocitrate
   and reduces NAD^+ to NADH. As shown in [103]Fig 4B and 4C, the
   activities of ICDH in C. glutamicum EGFP was lower than that in C.
   glutamicum BZH 001, indicating that the TCA cycle was weakened in C.
   glutamicum EGFP, a finding that was consistent with the result of the
   analysis of amino acids and organic acids. Glutamate dehydrogenase
   (GDH) catalyzes the conversion of α-ketoglutarate to Glu. We found that
   ICDH activity was lower but GDH enzyme activity was higher in C.
   glutamicum EGFP than in C. glutamicum BZH 001, a finding that was
   consistent with the content of Glu.

Analysis of the intracellular energy metabolism of C. glutamicum BZH 001 and
C. glutamicum EGFP

   As the most important energy source for metabolic pathways, ATP plays a
   vital role in cell growth and the production of the target products by
   affecting peptide folding, stress response, transportation, metabolic
   flux, and signal transduction [[104]33–[105]35]. A prior study had
   shown that approximately four ATP equivalents were required to form
   each polypeptide bond of GFP, and the amount of ATP needed from the
   intermediate metabolism to produce one unit of GFP was 333 [[106]35].
   Thus, the requirement of ATP is larger in C. glutamicum EGFP than in C.
   glutamicum BZH 001. In our transcriptome data, atpC, encoding the ATP
   synthase F0F1 subunit epsilon and NCgl1159, encoding the ATP synthase
   F0F1 subunit A, were up-regulated. As shown in [107]Fig 5, the ATP
   content of C. glutamicum EGFP was higher than that of C. glutamicum BZH
   001 during the entire fermentation process and increased rapidly since
   8 h, and the ATP concentration of both strains was reduced quickly
   after reaching a peak. It could be explained that as a sufficient
   supply of ATP was essential for the normal metabolism of cells and
   recombinant protein expression, the metabolic flux was changed to
   supply sufficient ATP for metabolism and protein synthesis.

Fig 5. Changes in the ATP content per mg protein of C. glutamicum BZH 001 and
C. glutamicum EGFP.

   [108]Fig 5
   [109]Open in a new tab

Discussion

   We cultivated the wild-type strain (C. glutamicum BZH001) and
   engineering strains (C. glutamicum EGFP) in a bioreactor to study the
   effect of EGFP expression on metabolism and identify the targets to
   improve the ability to express foreign proteins in C. glutamicum.
   Transcriptome analysis of the response of EGFP expression was carried
   out for 20 h, and DEG analysis was carried out through KEGG and GO
   enrichment. MVDA was also used to screen critical genes that play
   important roles in regulating EGFP expression, and then we detected the
   content of the intracellular and extracellular metabolites, including
   organic acids and amino acids, and the activities of the key enzymes in
   metabolic pathways. Furthermore, we detected the expression of key
   genes and ATP to study the changes in the material and energy
   metabolism in the fermentation process. Two effects of EGFP expression
   were found: 1. metabolism changes in EGFP expression; 2. the effect of
   EGFP expression on protein synthesis, secretion and expression
   regulation ([110]Fig 6, [111]S5 Table).

Fig 6. Metabolic shift of C. glutamicum EGFP under DO 30% concentration.

   [112]Fig 6
   [113]Open in a new tab

   The polygonal boxed enzymes and the hollow arrows indicate
   significantly up- or down-regulated enzyme activities, respectively.
   The square boxes represent the metabolites, and the solid arrows
   indicate the production of significantly up- or down-regulated
   metabolites, respectively. G-6-P: Glucose-6-Phosphate; 3-PG:
   3-Phosphoglycerate; PEP: Phosphoenolpyruvate; PK: Pyruvate kinase; PYR:
   Pyruvate; LDH: Lactate dehydrogenase; Lac: Lactate; PDH: Pyruvate
   dehydrogenase; AcCoA: Acetyl-CoA; PC: Pyruvic carboxylase; PEPCK:
   Phosphoenolpyruvate carboxykinase; OAA: Oxalacetic acid; CIT: Citrate;
   ICIT: Isocitrate; ICDH: Isocitrate dehydrogenase; α-KG:
   α-ketoglutarate; GDH: Glutamate dehydrogenase; GLT: Glutamate; Suc:
   Succinate; MAL: Malate; ICL: Isocitrate lyase; MS: Malate synthase.
   ABC: ABC transporter system. RP: ribosomal protein. Sec: Sec pathway.
   CSP: cold shock protein. TFs: transcriptional factors.

Metabolism changes for EGFP expression

   We found that there was a metabolism change under EGFP expression
   similar to oxygen limitation ([114]Fig 6). The cause may be that EGFP
   expression needs more ATP than the wild-type strain to satisfy the
   consumption of ATP—the metabolism changed to produce more ATP. First,
   the bacteria of C. glutamicum EGFP grew poorly compared with C.
   glutamicum BZH 001 under the 30% dissolved oxygen condition. The
   glycolysis pathway was up-regulated with glucose transfer and
   consumption increase, and more Gly, Lactate and acetate were
   accumulated under recombinant protein expression [[115]36]; however,
   the acetate accumulation had adverse effects on EGFP production owning
   to acetate overflowing is a manifestation of imbalance between glucose
   uptake and the limited activity of the TCA cycle and respiratory chain,
   which finally affect the macromolecule synthesis and ATP production
   [[116]30]. Meanwhile, because of the up-regulation of NCgl2248 encoding
   isocitrate lyase by 2.95-fold and the ICDH was lower in C. glutamicum
   EGFP than in C. glutamicum BZH 001, the glyoxylate cycle maybe
   increased. Second, the accumulation of pyruvic acid further leads to
   the synthesis of more metabolites, including lactic acid, valine and
   ILe. Third, because more acetate produced leading to carbon overflow of
   TCA, the increase of phosphoenolpyruvate (PEP) carboxylase (PPC)
   activity activated the anaplerotic reactions to supply more oxalacetic
   acid for the TCA cycle. The increase in oxalacetic acid caused more
   consumption of acetyl-CoA, further leading to decreased acetate
   accumulation, increased ATP and improved heterologous protein
   production [[117]37]. In addition, changes in the ATP content are
   closely related to redox, because of the regulation of each
   intracellular biological reaction process by intracellular energy
   levels and the influence of the energy and NADH synergy effect. In the
   metabolic process of C. glutamicum, substance metabolism is accompanied
   by ATP production and redox rebalance.

Effect of EGFP expression on protein synthesis, secretion and expression
regulation

   Although the expression of heterologous proteins caused expression
   changes of many key genes involved in the metabolic and energy pathways
   of C. glutamicum, changes in some genes related to protein synthesis
   and transport should be considered. First, the genes related to
   ribosomal protein were mostly down-regulated; however, NCgl2261
   encoding the 30S ribosomal protein S20 was up-regulated with EGFP
   expression. The cause may be that the down-regulation of ribosomal
   protein could save more ATP and substance. Second, the three genes of
   the sec pathway for proteins transported across the membrane was
   up-regulated; especially, NCgl1522 encoding secG was up-regulated by
   8.2-fold, and a previous study had shown that the overexpression of
   SecE and SecF could increase the extracellular amylase activity in C.
   glutamicum [[118]38]. Additionally, NCgl0303 encoding a cold shock
   protein belonging to the CspA family was up-regulated by 18.8-fold and
   may play an important role in the secretion of recombinant protein. A
   previous study showed that the absence of another protein of the CspA
   family—CspB—could increase the recombinant Fab secretion from C.
   glutamicum. Two genes are related to ABC transporter ATPase—NCgl0909
   and NCgl2060. The 31.2-fold up-regulation of NCgl0909 not only affected
   the activity of the ABC transporter system but also the concentration
   of ATP; thus, further investigation of this protein is warranted to
   improve the ATP levels. Third, Transcriptional regulation factor is
   another factor that impacts recombinant protein expression
   [[119]39,[120]40]. NCgl1504 is a DeoR family transcriptional regulator;
   the DeoR-type transcriptional regulator SugR acting as a repressor for
   genes of the PTS system in C. glutamicum was investigated [[121]41],
   showing that DeoR family genes could impact the recombinant protein
   expression when some genes of the PTS system are repressed. NCgl1844
   encodes RNA polymerase sigma factor SigB, whose disruption can increase
   the secretion of GFP and α-amylase by 3- to 5-fold compared with that
   in the wild type strain of C. glutamicum [[122]42]. However, in our
   transcription data, NCgl1844 showed slight up-regulation. We can
   down-regulate its expression in a further study for protein expression.

   The above results illustrated the influence of heterologous protein
   expression on metabolism, and many differentially expressed genes
   related to the metabolic flow changes were found, laying the foundation
   to identify genomic loci to change the flow of the host cell
   metabolism. Therefore, research concerning metabolism and the
   transcriptome regulation of heterologous protein expression in C.
   glutamicum has high academic value and broad application prospects.

Supporting information

   S1 Fig. The physical map of PXMJ19-EGFP.

   (PNG)
   [123]Click here for additional data file.^ (32.7KB, png)
   S2 Fig. The EGFP expression level of C. glutamicum EGFP in the
   fermentor.

   (JPG)
   [124]Click here for additional data file.^ (165.7KB, jpg)
   S1 Table. The specific primer used by real-time PCR.

   (DOCX)
   [125]Click here for additional data file.^ (15.4KB, docx)
   S2 Table. The correlation values and average CV among biological
   replicates.

   (DOC)
   [126]Click here for additional data file.^ (45KB, doc)
   S3 Table. Statistics of alignment analysis.

   (DOCX)
   [127]Click here for additional data file.^ (15.1KB, docx)
   S4 Table. Common critical genes from MVDA and DEG analysis of C.
   glutamicum EGFP compared with C. glutamicum BZH 001.

   (DOCX)
   [128]Click here for additional data file.^ (15.3KB, docx)
   S5 Table. Gene expression level of C. glutamicum EGFP compared to C.
   glutamicum BZH 001.

   (DOCX)
   [129]Click here for additional data file.^ (16.1KB, docx)

Acknowledgments