Sequence analysis of the native HGT1 from parent strain
There are two kinds of glucose transporters in A. niger, one is high-affinity and the other is low-affinity. HGT1 gene, as a high-affinity glucose transporter, was amplified from A. niger CGMCC 10142 genome. Then HGT1 sequence was compared with A. niger CBS 513.88 HGT1 sequences (GenBank: AM269996.1). The coding region totals 1644 bp and the coherence is 100%. The similarity of the amino acid sequences between A. niger HGT1 and MstA, LGT1, and A. nidulans MstE were 27.83%, 25.71%, 25.91%, respectively (shown in Figure 2B). Between HGT1 from A. niger and its homolog from Kluyveromyces lactis, the identity of amino acid sequence reached 40.87%, which was higher than with other glucose transporters. In this study, the structure of HGT1 from A. niger was predicted through homology modeling using Swiss-Prot (www.swissmodel.expasy.org/). The glucose transporter and D-xylose transporter were among the most similar related proteins (data not shown).
The phylogenetic tree of sugar (glucose) transporters from different fungi is shown in Figure 2A. At least several glucose transporters have been identified for both A. niger and A. nidulans. The HGT1 of A. niger CGMCC 10142 can be identified as the A. niger CBS 513.88 HGT1 (XP_001399197.1), which is located at the lower side of the phylogenetic tree. Several other glucose transporters that are close to HGT1 were XP_001394117.2 and XP_001390064.1. What’s more, The MstA in A. niger and most of the sugar transports identified in A. nidulans were located on the upper side of the phylogenetic tree.
It was found that in fermentations at relative lower glucose concentration (<50 g/L), the low-affinity glucose transporter is the limiting factor for glucose import into the cell [23]. The sugar concentration of the liquefied corn medium for CA fermentation is quite high at about 180 g/L and low-affinity glucose transporters can transport sugar fast at first. After 32 h fermentation, when the sugar concentration is declined quickly to less than 50 g/L, the low-affinity glucose transporters could not function, while the high-affinity glucose transporters could function even at the end of fermentation when the sugar concentration is extremely low. It’s weird that the high-affinity glucose transporters show extremely low-transcription level even no transcription. To improve the glucose utilization rate, the high-affinity glucose transporter HGT1 in A. niger CMCC 10142 was chosen to overexpress.
Construction of HGT1 overexpression cassettes and Screening of high CA production strains
The PglaA, Paox1, and HGT1 segments were obtained from the parent strain genome. Then the PglaA, HGT1, and Paox1, HGT1 was recombined into p80-HSVtk linear plasmid by MultiS kit (Vazyme Biotech Co., Ltd) to construct the new overexpression plasmids p20 and p21, respectively (shown in Figure 1B and Figure 1C). Both the plasmid p20 and p21 was successfully constructed by PCR and sequencing as showed in Figures S1a and S1b.
The ratios of the transparent halo zones due to acid production and colony diameter were calculated to select transformants with high CA production (Table S1). Then, genomic DNA was extracted from the transformants that showed a high CA production rate to confirm the correct insertion of the overexpression cassettes (promoter to HGT1) as shown in Figure S1c and S1d. Finally, eight transformants (A. niger 20-15, 20-16, 20-25, 20-27, 20-29, A. niger 21-8, 21-28 and 21-32) were selected for the ratios of the transparent halo zones (Figure 3B) and shake flask fermentation to screen higher transformants (Figure 3A).
Verification of genetic stability of the transformants by inserted expression cassettes on PCR
The inserted expression cassettes in the positive transformants were further verified. Firstly, The HGT1-hyg cassette of the transformants was verified as shown in Figure S1e. Then. the integration location was tested using primers of the ku70 total length. As shown in Figure S1f, except for p20-16 in the 8th lane, the ku70 original length (1800 bp) was amplified from all genomic DNA samples, and not the target fragment (5710 bp). These results indicated that the ku70 gene was partly destroyed in the strains 20-16. However, other transformants retained the intact ku70 gene, and the whole PglaA-HGT1-hyg segment was inserted. The overexpression frame from the upstream to the downstream homologous arm of ku70 was very long, totaling nearly 6000 bp. What’s more, it is common for filamentous fungi to lose heterologous genes, especially during long cultivation and passage. Therefore, the 8 strains were passaged for 15 generations to verify the transformants’ genetic stability. Then, conidia were prepared from the 8 identified stable transformants to verify their CA production ability in shake-flask fermentations.
Screening of higher CA production strains in shake-flask fermentations
The transformants (A. niger 20-15, 20-16, 20-25, 20-27, 20-29, A. niger 21-8, 21-28 and 21-32) were selected for the ratios of the transparent halo zones (Figure 3B) and shake flask fermentation to screen higher transformants (Figure 3A). As shown in Figure 3A and 3B, the CA production and the ratios of the transparent halo zones of all the eight transformants were higher than that of the A. niger CGMCC 10142. Two strains with the higher improvement of CA production were A. niger 20-15 and A. niger 21-8, with 15.8% and 12.8% comparing to the strain 10142. And the ratios of the transparent halo zones of A. niger 20-15 and A. niger 21-8 were 3.73 and 3.55 higher than strain 10142 with 2.85. The results indicated that the CA production ability of p20 transformants was higher than that of p21 transformants in shake flask fermentation. The higher production indicated that the PglaA promoter presents a higher transcription than the Paox1 promoter, leading to higher-level of HGT1 protein synthesis. Therefore, the two transformants A. niger 20-15 and 21-8 were chosen for further analysis by 30 L bioreactor-scale fermentation and real-time quantitative PCR.
CA fermentation and statistical analysis
As shown in Figure 4A, the morphology of the parental A. niger CGMCC 10142 strain as well as the transformants A. niger 20-15 and 21-8 for each sampling time are normal. The diameter of the mycelial pellets increased quickly from 8 h to 24 h, showing that the strains had entered the rapid growth phase. After 24 h the pellets of fungi showed a little increase and the colony began to sprout small mycelium, indicating that has begun to enter the period of CA production. Papagianni et al found reduction of mycelial clumps for the first 48 h of fermentation [35]. But in our research, the colony morphology didn’t change significantly from the period of CA production to the end of fermentation. Yet the mycelial pellets of A. niger 21-8 still showed some growth from 24 h to 32 h and maintained morphological stability to the end, too.
The CA production in the bioreactor-scale fermentations was also measured every 8 hours, as shown in Figure 4B. The CA production increased rapidly from 8 h, and the transformants showed a gradually improved CA fermentation ability. The CA production of A. niger 20-15 and 21-18 increased to 174.1 g/L and 169.4 g/L comparing to 10142 strain with 162.3 g/L which each improved by 7.3% and 4.4%, respectively, at the end of the fermentation. The productivity rates of each time stage of the fermentation process were calculated, as shown in Figure 4C. The highest productivity rates of A. niger 20-15 was reaching to 4.4 g/L/h which was higher than A. niger 21-8 and 10142 both with 4.1 g/L/h at 24 h of fermentation. The highest productivity rates of A. niger 21-8 was reaching to 4.2 g/L/h which was approximately equal to A. niger 20-15 but higher than A. niger CGMCC 10142 with 4.0 g/L/h at 32 h, and maintained higher level at 40 h. This indicated that the 24-h CA metabolism level of A. niger 20-15 was higher than that of A. niger 21-8 and A. niger CGMCC 10142, probably due to the fact that 21-8 did not have such a strong HGT1 overexpression burden. The biggest differences in productivity rate with the starting strains were among 56 h-64 h at the end of fermentation. Among 56 h-64 h, the CA productivity rates of A. niger 20-15 and A. niger 21-8 were 1.6-0.6 g/L/h and 1.0-0.7 g/L/h, respectively, which were 2.1-1.5 and 1.3-1.8 times higher than the A. niger CGMCC 10142 with 0.7-0.4 g/L/h. The above results further confirmed that the overexpression of HGT1 promoted CA production at the later stage of fermentation when the nutrition were running out.
During the fermentation process, the total sugar in the fermentation broth was measured every 8 hours. The results of residual total sugar at 48 h, 56 h and 64 h were 36.6 g/L, 21.7 g/L, 10.6 g/L of A. niger 20-15 which were higher than A. niger 21-8 with 33.8 g/L, 18.9 g/L and 8.8 g/L but lower than A. niger CGMCC 10142 with 37.4 g/L, 22.9 g/L and 12.7 g/L, respectively (shown in Figure 5A). This indicated that the overexpression of HGT1 had a positive influence in the transformant strains at 48 h, 56 h and 64 h of the fermentation. This result was consistent with an earlier study [36]. What’s more, HGT1 did not function effectively because of the high sugar concentration (data not shown). Earlier studies found that low-affinity glucose transporter is formed when the glucose concentration is 15% and is highly expressed when it drops to more than 8% [23, 24]. Our results also showed that the most significant changes were observed when the fermentation 32 h sugar concentration dropped below 8% HGT1 began to play a more important role. At the end of fermentation (64 h), the total residual sugar of A. niger 20-15 and A. niger 21-8 decreased 16.5% and 30.7% compared with A. niger CGMCC 10142. This result indicated that the overexpression of HGT1 significantly improved the sugar consumption rate.
A proportion of the initial sugar is lumping which was unutilized by microorganisms during CA fermentation leading to waste of raw materials. Therefore, the total reducing sugar concentrations in the fermentation broth were measured at 48 h, 56 h, and 64 h as shown in Figure 5B. The total reducing sugar concentrations at 48 h, 56 h, and 64 h of A. niger 20-15 were 16.0 g/L, 5.9 g/L, and 2.1 g/L which were less than A. niger 21-8 with 18.1 g/L, 7.4 g/L, and 2.7 g/L. Both of them were less than A. niger CGMCC 10142 with 19.1 g/L, 8.3 g/L and 3.8 g/L, and the utilization rates of the total reducing sugar of A. niger 20-15 and 21-8 were improved 44.7% and 26.3% comparing to A. niger CGMCC 10142 at the end of fermentation. What’s more, we found that the reducing sugar concentrations were increased rapidly during the first 16 h, and especially during the first 8 h. It then decreased in the subsequent fermentation period with increasing CA production, during which a large amount of glucose was used to synthesize CA. Consequently, the reducing sugar concentration decreased rapidly after 24 h (data not shown). Therefore, the overexpressed-HGT1 transformants showed the higher glucose utilization ability and also improved CA production.
Meanwhile, we calculated the glucose-CA conversation rate based on the measured values of total sugar consumption and CA production, as shown in Figure 5C. As expected, the overexpressed-HGT1 strains showed higher conversation rates than the original strain. The final conversation rates of A. niger 20-15 and 21-8 were 102.4% and 100.4%, respectively. The CA fermentation performance was in the order A. niger 20-15>A. niger 21-8>A. niger CGMCC 10142. It is worth noting that the overexpression strain with the PglaA promoter showed the best performance.
Real-time quantitative PCR
The relative transcription levels of the HGT1 gene (XM_001399160.2), citrate synthase (CS) gene (XM_001393946), and glucokinase gene (XM_001395875.2) during the fermentation process were measured to explore whether the overexpression of HGT1 influences the metabolism and accumulation of CA in A. niger. Mycelial pellet simples of A. niger 20-15 and 21-8 were taken at 12 h and 48 h, separated from the culture broth through filtration, and the total RNA was extracted for real-time quantitative PCR.
The relative expression of HGT1 is shown in Figure 6A. The expression level of HGT1 in the overexpression strains was much higher than in the control at 12 h and 48 h. The overexpression of HGT1 in A. niger 20-15 was about 219 times and 208 times higher than the original strain at 12 h and 48 h, respectively, which indicated that the glucose utilization efficiency of the HGT1 overexpression strain was also higher than that of the A. niger CGMCC 10142. What’ s more, the HGT1 overexpression strain with the PglaA promoter showed 2.5 times higher expression than the strain with Paox1 at 12 h, as well as 5 times higher at 48 h. PglaA is a widely used strong fungal promoter that can be induced by starch or dextrose [37]. Paox1 is the promoter of alternative oxidase in A. niger [8]. The overexpression of the low-affinity glucose transporter increased the substrate uptake at high glucose concentrations, partly releasing the limit of glucose consumption in the strains. The HGT1 transporter accelerated sugar transport at the end of CA fermentation when the remaining glucose concentration was low. This could reduce the final total residual sugar and improve the economic value of the fermentation.
Notably, the relative expression levels of glucokinase in the HGT1 overexpression strains were higher than in the parent strain, especially at 48 h (Figure 6B). The glucokinase expression of A. niger 20-15 was 8.4 times higher than in the parent strain and that of A. niger 20-8 was 1.4 times higher at 48 h, which indicated that the overexpression strains had a higher glucose utilization rate and improved CA producing ability. The results indicated that the overexpression of HGT1 increased the glucose uptake efficiency, which provided adequate carbon flux for increased CA production in the transformants. Glucokinase plays a key role in the metabolic activation of glucose to 6-p-glucose, and its activity is controlled via feedback inhibition by 6-p-glucose and ADP. The increased glucokinase activity could therefore directly improve the CA production at the end of fermentation process.
As shown in Figure 6C, the expression of citrate synthetase (CS) presented a similar trend to glucokinase. The CS expression of the engineered A. niger 20-15 was 6.4 times higher than that of the parent strain at 48 h. CS is necessary for the synthesis of CA as indicated by its name, and therefore plays a key role in CA metabolism. When the CS gene in A. niger is deleted, CA production falls to almost zero. High expression of HGT1 therefore increased the expression of CS and contributed to abundant CA accumulation. The changes in the relative expression of CS indicated that high carbon flux increased the CA yields of the transformants. It is therefore possible that substrate limitation is one of the reasons for low enzyme activity in the parent strain.
In this study, we obtained some strains which overexpress the glucose transporter HGT1 using two promoters PglaA and Paox1. The results of real-time quantitative PCR indicated that PglaA is a stronger promoter than Paox1. PglaA is the native promoter of glucoamylase in A. niger, and it is widely used as a strong promoter. By contrast Paox1 is the native promoter of alternative oxidase in the alternative respiratory chain. Our past research found that the overexpression of aox1 also could effectively increase CA production [8]. Therefore, these two promoters were used to induce HGT1 overexpression in the present study. It is reasonable that the PglaA-induced strain A. niger 20-15 had higher HGT1 transciption level than the strain using Paox1 as shown in Figure 6A.
Notably, although PglaA is a stronger promoter than Paox1, there were 1.5 times higher transcription levels of glucokinase and CS induced by Paox1 than PglaA at 12 h where the strains were still in the active growth phase (shown in Figure 6B and 6C). When mycelium grows vigorously at 12 hours, dissolved oxygen and nutrition may become limiting factors. In this case, Paox1 activates the expression of the key enzyme Aox1 in the secondary respiratory chain that does not require high amounts of oxygen, which may increase the metabolic flow, thereby increasing the expression of glucokinase and citrate synthase [6]. Interestingly, at 48 h of the fermentation time in which the carbon source becomes limited, Paox1 cannot influence the secondary respiratory chain. Therefore, the PglaA-containing transformant A. niger 20-15 showed higher transcription levels of glucokinase and CS as shown in Figure 6B and 6C.
L. Wang et al. found that when a native promoter of the original strain is used in an engineered strain, could not only drive the expression of the target genes, but it may also increase the expression of the original gene of the promoter itself [38]. So we speculated that Paox1 may not only drive the expression of the HGT1 gene but also activate its own aox1 gene to some extent and thereby indirectly increase the fermentation performance observed in A. niger 21-8. In view of this, it may be useful to verify the relative expression of aox1 and related enzymes in further studies. What’s more, insertion locus or copy numbers of the overexpression cassettes may also influence the expression levels. The genes that are important for CA synthesis, such as CS, and glucokinase were expressed at higher levels in the HGT1 overexpression strains than in the A. niger CGMCC 10142. It is possible that sufficient glucose increased the expression of these genes and thereby improved the CA production performance of A. niger [23]. Besides, the strain used in this study is already an industrial CA production strain, and its metabolite levels are higher than in undomesticated wild-type strains.
Glucose transport is the first step from sugar to CA, but it does not influence the CA fermentation directly. In this study, we found that the overexpression of HGT1 improved the CA fermentation performance in terms of reducing sugar consumption and total residual sugar in the fermentation process. Furthermore, the final glucose-CA conversation rates of A. niger 20-15 and 21-8 both increased to 102.4% and 100.4%, which nearly to theoretical glucose-CA conversion rate of just 106.7% when all the resources are used to produce CA and none for growth [25]. Finally, the CA production of the engineered strains also increased by 7.3% for A. niger 20-15 and 4.4% for A. niger 21-8. Therefore, sugar intake may still be a bottleneck in the fermentation process.