Engineering cofactor metabolism for improved protein and glucoamylase production in Aspergillus niger [original]

Suietal. Microb Cell Fact (2020) 19:198

https://doi.org/10.1186/s12934-020-01450-w

RESEARCH

Engineering cofactor metabolism

forimproved protein andglucoamylase

production inAspergillus niger

Yu‑fei Sui1,2, Tabea Schütze2 , Li‑ming Ouyang1, Hongzhong Lu3, Peng Liu1, Xianzun Xiao1, Jie Qi1,

Ying‑Ping Zhuang1* and Vera Meyer2*

Abstract

Background: Nicotinamide adenine dinucleotide phosphate (NADPH) is an important cofactor ensuring intracel‑

lular redox balance, anabolism and cell growth in all living systems. Our recent multi‑omics analyses of glucoamylase

(GlaA) biosynthesis in the filamentous fungal cell factory Aspergillus niger indicated that low availability of NADPH

might be a limiting factor for GlaA overproduction.

Results: We thus employed the Design‑Build‑Test‑Learn cycle for metabolic engineering to identify and prioritize

effective cofactor engineering strategies for GlaA overproduction. Based on available metabolomics and 13C meta‑

bolic flux analysis data, we individually overexpressed seven predicted genes encoding NADPH generation enzymes

under the control of the Tet‑on gene switch in two A. niger recipient strains, one carrying a single and one carrying

seven glaA gene copies, respectively, to test their individual effects on GlaA and total protein overproduction. Both

strains were selected to understand if a strong pull towards glaA biosynthesis (seven gene copies) mandates a higher

NADPH supply compared to the native condition (one gene copy). Detailed analysis of all 14 strains cultivated in

shake flask cultures uncovered that overexpression of the gsdA gene (glucose 6‑phosphate dehydrogenase), gndA

gene (6‑phosphogluconate dehydrogenase) and maeA gene (NADP‑dependent malic enzyme) supported GlaA

production on a subtle (10%) but significant level in the background strain carrying seven glaA gene copies. We thus

performed maltose‑limited chemostat cultures combining metabolome analysis for these three isolates to charac‑

terize metabolic‑level fluctuations caused by cofactor engineering. In these cultures, overexpression of either the

gndA or maeA gene increased the intracellular NADPH pool by 45% and 66%, and the yield of GlaA by 65% and 30%,

respectively. In contrast, overexpression of the gsdA gene had a negative effect on both total protein and glucoamyl‑

ase production.

Conclusions: This data suggests for the first time that increased NADPH availability can indeed underpin protein

and especially GlaA production in strains where a strong pull towards GlaA biosynthesis exists. This data also indicates

that the highest impact on GlaA production can be engineered on a genetic level by increasing the flux through

the pentose phosphate pathway (gndA gene) followed by engineering the flux through the reverse TCA cycle (maeA

gene). We thus propose that NADPH cofactor engineering is indeed a valid strategy for metabolic engineering of A.

adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and

the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material

in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material

is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the

permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco

mmons .org/licen ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/publi cdoma in/

zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Open Access

Microbial Cell Factories

*Correspondence: [email protected].cn; vera.meyer@tu‑berlin.de

1 State Key Laboratory of Bioreactor Engineering, East China University

of Science and Technology, Shanghai 200237, People’s Republic of China

2 Chair of Applied and Molecular Microbiology, Institute of Biotechnology,

Technische Universität Berlin, Straße des 17. Juni 135, 10623 Berlin,

Germany

Full list of author information is available at the end of the article

Page 2 of 17

Suietal. Microb Cell Fact (2020) 19:198

Background

The filamentous fungus Aspergillus niger is one of the

main cell factories used nowadays in the industry for

homologous or heterologous protein production due

to its extraordinary ability for protein expression and

secretion [1–3]. The Design-Build-Test-Learn (DBTL)

cycle is an increasingly adopted systematic meta-

bolic engineering strategy to achieve the desired out-

come through reconstructing heterologous metabolic

pathways or rewiring native metabolic activities [4,

5]. Rational strain development of cell factories can

be improved by the iterative application of the DBTL

cycles, which not only contributes to the optimization of

biomanufacturing processes, it is also advantageous to

build a complete metabolic model of engineered cells to

deepen our understanding of cellular metabolism. Note-

worthy, the advance of genetic engineering has speeded

up the DBTL cycle of metabolic engineering [6]. For the

cell factory A. niger, several genetic approaches have

proven their potency to improve its enzyme producing

capability, including protein carrier approaches, tun-

able Tet-on driven gene expression, and morphology

engineering, to name but a few [1, 7–9]. However, the

impact of cofactor engineering, i.e., the rebalance of the

intracellular redox status, on protein production has not

been systematically studied in A. niger.

NADPH is a limiting factor for the biosynthesis of

amino acids that are the building blocks of proteins. For

instance, 3mol and 4mol of NADPH is required for pro-

ducing 1 mol of arginine and lysine, respectively [10].

Thus, adequate cytosolic NADPH supply is indispensa-

ble to maintain the intracellular redox balance and serves

as a driving force for efficient amino acid biosynthesis

[11]. NADPH also provides the main anabolic reducing

power for biomass growth, lipid formation, and also for

natural product biosynthesis [12]. Indeed, cofactor engi-

neering has been reported to improve productivities in

the bacterial cell factories Escherichia coli [13, 14], and

Corynebacterium glutamicum, as well as in the yeast cell

factory Yarrowia lipolytica [15]. Two common strategies

have mainly been employed to optimize the availabil-

ity of NADPH. One is to activate the enzyme activities

of NAD(H) kinases (EC 2.7.1.86, EC 2.7.1.23) which are

used to obtain NADPH or NADP + through phospho-

rylation of NADH and NAD + , respectively. The other is

to modulate the expression strength of typical NADPH

generating enzymes of the glycolytic pathway, the pen-

tose phosphate pathway or the citric acid cycle. These

include glucose-6-phosphate dehydrogenase (G6PDH),

6-phospho-gluconate dehydrogenase (6PGDH), NADP-

dependent isocitrate dehydrogenase (NADP-ICDH),

and NADP-dependent malic enzyme (NADP-ME) [16,

17]. Notably, heterologous protein expression in Pichia

pastoris and A. niger can be triggered through boosted

carbon flux to the pentose phosphate pathway (PPP), a

catabolic pathway also known to produce NADPH [18,

19]. This suggests that central carbon metabolism may

have evolved to ensure the production of cellular compo-

nents under the balance of energy production and con-

sumption [4]. In agreement, the metabolic flux through

the PPP increased by 15–26% compared to the parental

strains when GlaA was overproduced in A. niger [20] or

the enzyme amylase overproduced in A. oryzae [21].

In the past two decades, extensive studies have focused

on engineering a high flux through the PPP in E. coli,

C. glutamicum, A. nidulans, and A. niger [11, 22–25].

A block of the glycolytic pathway by down-regulating

the pgi gene encoding a phosphoglucose isomerase was

one successful strategy in C. glutamicum [11]. In order

to elevate the NADPH pool originating from the PPP

in A. niger, the gsdA gene (glucose 6-phosphate dehy-

drogenase), the gndA gene (6-phosphogluconate dehy-

drogenase) and the tktA gene (transketolase) were

individually overexpressed in A. niger. Strong overexpres-

sion of gndA led to a nine-fold increase in intracellular

NADPH concentration, while gsdA and tktA affected the

NADPH level only weakly [25]. However, any correlation

between the NADPH supply and enzyme overproduction

remained unclear.

Irrespective of the importance of the PPP for NADPH

regeneration, an efficient carbon economy is only guaran-

teed when the carbon flux enters the glycolytic pathway

(Embden-Meyerhoff-Parnass pathway, EMP) instead of the

PPP because the PPP releases one carbon as CO2 when oxi-

dizing 1mol of hexose. Takeno etal. [26] thus substituted

the endogenous NAD-dependent glyceraldehyde-3-phos-

phate dehydrogenase (GAPDH) in the EMP in C. glu-

tamicum with a heterologous NADP-dependent GAPDH,

leading to 2mol of NADPH generation instead of 2mol

NADH from 1mol of hexose. This genetic modification

niger to improve GlaA production, a strategy which is certainly also applicable to the rational design of other microbial

cell factories.

Keywords: Aspergillus niger, NADPH, Genetic engineering, CRISPR/Cas9, Tet‑on, Metabolic engineering, Chemostat,

Glucoamylase

Page 3 of 17

Suietal. Microb Cell Fact (2020) 19:198

provoked a substantial improvement in the yield of L-lysine

production by 70–120%. Similar strategies also have been

followed to overproduce ethanol in the yeast Saccharomy-

ces cerevisiae [27] or lycopene and ε-caprolactone in the

bacterium Clostridium acetobutylicum [28]. Likewise, cyto-

solic NADP-ME has been shown to positively affect lipid

accumulation in oleaginous fungi [29, 30].

As summarized above, a wealth of metabolomic and

fluxomic data in A. niger demonstrated that strains

adapted to protein overproducing conditions channel a

higher carbon flux through the PPP. However, cofactor

engineering has not been considered yet or performed

in A. niger to guide enzyme overproduction. We thus

mined our recently published genome-scale metabolic

network model (GSMM) developed for the A. niger pro-

tein producing reference strain CBS 513.88 [31]. This

iHL1210 model identified the involvement of NADPH

in 173 intracellular redox reactions in A. niger, includ-

ing 49 NADPH generating reactions [31]. Notably, the

GSMM did not predict any NADPH/NADP + shuttle in

the mitochondrial membrane, and we thus concluded

that any mitochondrial NADPH is unlikely to become

directly consumed by cytosolic amino acid biosynthe-

sis. Overall, the GSMM predicted that seven potential

NADPH generating enzymes are of importance for GlaA

production in A. niger (Table1, Fig.1): two enzymes of

the cytosolic PPP (glucose-6-phosphate dehydrogenase,

G6PDH; 6-phosphogluconate dehydrogenase 6PGDH),

two cytosolic NADP-dependent enzymes (NADP-ICDH

and NADP-ME) and three uncharacterized open read-

ing frames (An12g04590, An14g00430, An16g02510).

An12g04590, An14g00430 show high homology to

NADP + oxidoreductases, and An16g02510 displays

homology to alcohol dehydrogenases. In order to evalu-

ate whether the model prediction is strain-dependent,

we individually overexpressed all seven candidate genes

in two A. niger host strains. Strain AB4.1 produces native

levels of GlaA as it carries one glaA gene copy, and strain

B36 is a derivative thereof, carrying seven glaA gene

copies and is thus a high-yield GlaA producing strain

[32]. All 14 strains were first investigated in shake flask-

level cultivations. Based on the data gained, three engi-

neered strains were selected for chemostat cultivations

to decipher the association among genetic perturbation,

NADPH availability, and GlaA production in A. niger.

Results

Strain generation using CRISPR/Cas9 technology

andthesynthetic Tet‑on gene switch

In order to compare the effect of the seven selected genes

on GlaA production in an A. niger strain carrying one

glaA (AB4.1) or seven glaA (B36) gene copies, we first

had to ensure that the introduced genetic modifications

would allow us to directly compare the observed phe-

notypes. This required that the introduced genes would

be under the same genetic control and furthermore

introduced at the same genomic locus in both recipient

strains. We thus decided to integrate an additional copy

of all candidate genes under the control of the strong

and tunable Tet-on gene switch into the pyrG locus of

A. niger. This gene switch is inducible by the addition of

doxycycline (DOX) to the culture medium, is tight in the

absence of DOX and metabolism-independent in A. niger

[33]. It has furthermore been shown to strongly induce

gene expression up to levels above the glucoamylase

gene, which is one of the highest expressed genes in A.

niger [33–35].

Strain AB4.1 is a uridine-auxotroph due to a defective

pyrG gene, a locus that is perfectly suited for gene tar-

geting and screening purposes. The introduction of an

Table 1 GSMM-predicted NADPH producing reactions inA. niger

[m] reactions in the mitochondrion; Rn, reaction. All reactions listed here were predicted in Lu etal. [31]

name Rn description Formula Gene

R25 Glucose 6‑phosphate‑dehydrogenase (gsdA) G6P + NADP = > D6PGL + NADPH + H An02g12140

R27 Phosphogluconate dehydrogenase (gndA) 6PGC + NADP = > Ru5P + CO2 + NADP An11g02040

R36 Isocitrate dehydrogenase (icdA) (NADP +) ICIT[m] + NADP[m] = > AKG[m] + CO2[m] + NADPH[m] An02g12430

R38 Isocitrate dehydrogenase (NADP +) ICIT + NADP = > AKG + CO2 + NADPH An02g12430

R55 Malic enzyme

(NADP‑specific) (maeA)MAL + NADP = > PYR + CO2 + NADPH An05g00930

R57 Malic enzyme

(NADP‑specific) MAL[m] + NADP [m] = > PYR

[m] + CO2[m] + NADPH [m] An05g00930

R110 (S)‑3‑Hydroxybutanoyl‑CoA: NADP + oxidoreductase 3HBCoA[m] + NADP [m] < = > AACCoA[m] + NADPH [m] +H[m] An14g00430

R125 dihydrofolate:NADP + oxidoreductase NADP [m] + DHF[m] < = > NADPH [m] + FOLATE[m] An12g04590

R190 Alcohol dehydrogenase ETH + NADP < = > ACAL + NADPH + H An16g02510

Page 4 of 17

Suietal. Microb Cell Fact (2020) 19:198

intact pyrG copy at this locus occurs efficiently and con-

fers uridine prototrophy in A. niger [36]. However, strain

B36 does carry an intact pyrG gene [37]. We thus first

mutated the pyrG locus in this strain in order to apply the

same gene targeting strategy for all seven genes in both

recipient strains. We edited the pyrG gene in B36 by fol-

lowing a CRISPR/Cas9 strategy that employed ribonu-

cleoprotein particles. This approach was first published

for the penicillin producer Penicillium chrysogenum [38]

and has later been successfully established in other fun-

gal cell factories [39]. As explained in detail in Additional

file1: Fig. S1, this approach enabled us to obtain a deriva-

tive of B36, strain YS20.2, which carries a 195bp deletion

within the pyrG ORF and is therefore unable to grow on

medium lacking uridine or uracil. Both recipient strains,

AB4.1 and YS20.2, were eventually used to integrate Tet-

on driven candidate genes at the pyrG locus (for details

see Materials and Methods and Additional file 1: Fig.

S2). Respective genetic modifications were proven by

PCR and Southern blot analyses (Additional file1: Figs.

S3, S5, S6, S7). All 14 strains obtained are summarized in

Table3. We finally also decided to delete the native ORFs

of An14g00430 and An16g02510 in their respective Tet-

on driven overexpression strains in order to analyze their

deletion phenotypes.

The impact ofNADPH engineering onGlaA production

isstrain‑dependent

All 14 strains were subjected to batch cultivations

in shake-flask format, whereby a medium contain-

ing maltose as GlaA-inducing carbon source was

used. FW35.1 (a pyrG + derivative of AB4.1) and B36

were taken along as corresponding reference strains.

DOX-induced gene expression in all seven AB4.1

derivatives was about 1.5–2.7 times higher compared

to the reference strain FW35.1 as examined by qRT-

PCR (Additional file1: Fig. S8, TableS4). Although

this led to an elevated NADPH pool of about 30% in

Fig. 1 Pathway map highlighting all seven genes modified during this study in red. The cytosolic glycolytic pathway, the pentose phosphate

pathway and the mitochondrially located citric acid cycle are shown

Page 5 of 17

Suietal. Microb Cell Fact (2020) 19:198

the case of gndA, icdA or An16g02510, no significant

increase in GlaA enzyme activity was observed for all

of seven strains compared to the FW35.1 reference

(Additional file1: Fig. S8, TableS4). However, when

all seven candidate genes were overexpressed in the

YS20.2 background strain containing seven glaA gene

copies, increased transcript levels were similar as in

AB4.1, but for An16g02510 higher transcription levels

(fourfold) were observed (Fig.2b). Noteworthy, over-

expression of gndA displayed the highest effect on the

Fig. 2 Data for shake flask‑level cultivations of all engineered strains in the YS20.2 background in relation to the control strain B36. a Dry cell weight

(DCW); b Relative expression level of glaA and engineered genes; c Total secreted protein per gram biomass at 72 h after inoculation; d Enzyme

activity of GlaA per gram biomass at 72 h after inoculation; e Intracellular NADPH concentration in the exponential phase; f Comparison between

engineered strains in the AB4.1 and YS20.2 background, respectively. All experiments were conducted in biological quadruplicates. Significance

values were calculated with the two‑tailed t‑test with independent variables (*p < 0.05, **p < 0.01, ***p < 0.001)

Loading more pages...