Assembly of nitrogenase biosynthetic pathway in Saccharomyces cerevisiae by using polyprotein strategy

Nitrogenase in some bacteria and archaea catalyzes conversion of N2 to ammonia. To reconstitute a nitrogenase biosynthetic pathway in a eukaryotic host is still a challenge, since synthesis of nitrogenase requires a large number of nif (nitrogen fixation) genes. Viral 2A peptide mediated “cleavage” of polyprotein is one of strategies for multigene co-expression. Here, we show that cleavage efficiency of NifB-2A-NifH polyprotein linked by four different 2A peptides (P2A, T2A, E2A, and F2A) in Saccharomyces cerevisiae ranges from ~50% to ~90%. The presence of a 2A tail in NifB, NifH, and NifD does not affect their activity. Western blotting shows that 9 Nif proteins (NifB, NifH, NifD, NifK, NifE, NifN, NifX, HesA, and NifV) from Paenibacillus polymyxa that are fused into two polyproteins via 2A peptides are co-expressed in S. cerevisiae. Expressed NifH from Klebsiella oxytoca NifU and NifS and P. polymyxa NifH fusion linked via 2A in S. cerevisiae exhibits Fe protein activity.


Introduction
Although nitrogen is abundant in the atmosphere as N 2 on earth, the N 2 is not easily available to living organisms, and the bio-available nitrogen (such as ammonia and nitrogen oxides) limits the productivity of major crops (Hoffman et al., 2014). The addition of industrially produced nitrogenous fertilizer can maintain crop productivity, but overuse of chemical nitrogen fertilizers leads to economic costs and environmental pollution (Zhang et al., 2015). One way to reduce use of fertilizers is to transfer biological nitrogen fixation to cereal crops that can fix nitrogen (Beatty and Good, 2011;Curatti and Rubio, 2014;Oldroyd and Dixon, 2014;Vicente and Dean, 2017;Good, 2018;Li and Chen, 2020).
Biological nitrogen fixation, mainly carried out by Mo nitrogenase enzyme in some bacteria and archaea, is a process in which atmospheric N 2 is converted to ammonia (Dos et al., 2012;He et al., 2022). Nitrogenase consists of two metalloprotein components: Fe protein and MoFe protein (Georgiadis et al., 1992). Fe protein, also termed as NifH protein, is a homodimer (encoded by nifH) bridged by an intersubunit  cluster that serves as the obligate electron donor to the MoFe protein. MoFe protein, also termed as NifDK protein, is a heterotetramer (encoded by nifD and nifK) that contains two metalloclusters: FeMo-cofactor (Mo − 7Fe−9S − C − homocitrate) that serves as the active site of substrate binding and reduction, and the P cluster (8Fe−7S) that shuttles electrons to FeMo-co (Burén et al., 2020). Apart from nifH, nifD and nifK encode the structural subunits of nitrogenase, several genes are Biological nitrogen fixation requires a large number of nif genes. How to engineer large numbers of nif genes into eukaryotic organism is still a challenge. Co-expression of multiple genes at a desired ratio is one strategy to solve the challenge. 2A peptides, the "self-cleaving" small (18-22 amino acids) peptides, are used to express multiple proteins from a single open reading frame (ORF) in eukaryotic cells (Ryan et al., 1999). When two genes are linked via a 2A peptide sequence, two separate proteins are generated from an ORF through a novel "cleavage" event within the 2A peptide sequence (Atkins et al., 2007;Doronina et al., 2008). Cleavage occurs at the end of the 2A peptide sequence through translation by 80S ribosomes (Szymczak-Workman et al., 2012). The cleavage of 2A peptides is not associated with the host cell proteases or viral proteinases, instead, associated with the highly conserved C-terminal of the 2A peptides (-DxExNPGP-) (Palmenberg et al., 1992;Donnelly et al., 2001;Sharma et al., 2012). 2A peptides were first discovered in the foot-andmouth disease virus (FMDV) (Ryan and Drew, 1994). Since then, many 2A-like sequences have been identified in other viruses (Szymczak and Vignali, 2005;Wang and Marchisio, 2021 In this study, cleavage efficiency of NifB and NifH proteins linked via four different 2A peptides in S. cerevisiae are investigated by using Western blotting and yeast two-hybrid system analysis. Then, nine Nif proteins (NifB, NifH, NifD, NifK, NifE, NifN, NifX, HesA, and NifV) from P. polymyxa WLY78 are fused into two huge proteins and then they are co-expressed in S. cerevisiae. P. polymyxa NifH and K. oxytoca NifS and NifU are fused into a polyprotein and co-expressed in S. cerevisiae, and the expressed NifH from the huge protein (NifS-2A-NifU-2A-NifH) exhibits Fe activity. This study shows the potential of utilization of 2A peptides to express multicomponent nitrogenase system in S. cerevisiae.

Results and discussion
Cleavage efficiency of NifB and NifH proteins linked via 2A peptide in yeast To test whether 2A peptide was able of separating bacterial nitrogenase proteins, two Nif proteins (NifB and NifH), which are the first and second proteins in Nif-cluster (NifBHDKENXHesANifV) of P. polymyxa WLY78, were chosen to be tested (Wang et al., 2013). The nifB and nifH genes, codon optimized for S. cerevisiae, were linked into a single open reading frame (ORF) encoding a polyprotein (designated NifB-2A-NifH) by four different 2A peptides (P2A, T2A, E2A and F2A), respectively ( Figure 1A). Then, each of a single ORF was cloned into pRS423-GAL1p plasmid under the control of GAL1 promoter ( Figure 1B), and each plasmid was transformed into S. cerevisiae YSG50. To determine the cleavage efficiency of the 2A peptide linked NifB and NifH, total protein extracts were prepared from aerobically grown yeast cultures and analyzed by Western blotting with the antibody against the NifH and NifB proteins of P. polymyxa WLY78 which was expressed and purified from E. coli BL21. The calculated molecular weights of NifH and NifB were 31.5 KDa and 55.0 KDa, respectively. The NifH protein was detectable in Frontiers in Microbiology 03 frontiersin.org the recombinant S. cerevisiae strains but not in the wild-type S. cerevisiae YSG50 ( Figure 1C). All the four 2A peptides evaluated showed cleavage with different efficiency, ranging from ~50% to ~90%. The best performance was achieved using P2A and T2A, in which the cleaved form represented more than 80% ( Figure 1D). Our data are consistent with the reports that order of the work efficiency is P2A, T2A, F2A, and E2A (Szymczak-Workman et al., 2012). The results indicate that P2A and T2A should be preferred for expression of Nif proteins in yeast. In addition, the different cleavage efficiency among the four 2A peptides was verified by yeast two hybrid assay in vivo. Two vectors pGBKT7 carrying GAL4 DNA-binding domain (BD) and pGADT7 carrying a GAL4 activation domain (AD) are usually used in yeast two hybrid (Supplementary Figure S1A). Here, a middle vector pGBKT7-AD that simultaneously carries AD and BD of GAL4 was constructed by cloning the GAL4 activation domain (AD) to vector pGBKT7 as described in Materials and Methods. Then, five nifB and nifH fusions linked with four different 2A peptides or without 2A peptide were individually in frame inserted between BD and AD domains of GAL4 in vector pGBKT7-AD, and then GAL4 BD, nifB, nifH and GLA4 AD were fused into a single ORF under control of ADH1 promoter (Supplementary Figure S1B). These plasmids were individually transformed into S. cerevisiae YSG50, yielding five recombinant strains: S. cerevisiae Y2H-BH carrying NifB-NifH, S. cerevisiae Y2H-BP2AH carrying NifB-P2A-NifH, S. cerevisiae Y2H-BT2AH carrying NifBT2ANifH, S. cerevisiae Y2H-BE2AH carrying NifH-E2A-NifH and S. cerevisiae Y2H-BF2AH carrying NifB-F2A-NifH (Supplementary Table S1). Yeast strain Y2H Gold carrying vector pGBKT7 was used as a negative control. All yeast strains grew well on the SC-Trp plates (Supplementary Figure S1C), indicating that TRP1 carried in vector pGBKT7 was expressed. When grown on SC-Trp-His-Ade plates, S. cerevisiae BH carrying NifB-NifH grew well, suggesting that GAL4 AD-NifB-NifH-GAL4 BD was a polyprotein. Whereas, S. cerevisiae BP2AH carrying NifB-P2A-NifH, S. cerevisiae BT2AH carrying NifBT2ANifH, S. cerevisiae BE2AH carrying NifH-E2A-NifH could not grow on SC-Trp-His-Ade plates, suggesting that NifB and NifH was effectively separated by P2A, T2A and E2A. However, S. cerevisiae BE2AH carrying NifH-F2A-NifH could grow well as S. cerevisiae BH carrying NifB-NifH did on SC-Trp-His-Ade plates, suggesting that E2A could not effectively separate NifB and NifH. The results are consistent with the western blotting data described above.

Tailing-tolerance assay
Since the 2A tail remains attached to the carboxyl terminus of the upstream protein after cleavage, it necessary to investigate influence   Figure S2). The data suggest that the 2A tails did not affect activity of NifH protein.
Further, we selected P2A, an efficient cutter, to verify its influence on the three major nitrogenase proteins NifB, NifD, and NifK. As described in Materials and Methods, complementation strains P. polymyxa Btail, P. polymyxa Dtail and P. polymyxa Ktail were constructed by using nifB, nifD and nifK carrying P2A tail to complement ΔnifB, ΔnifD, ΔnifK mutants of P. polymyxa, respectively. P. polymyxa proB, P. polymyxa proD and P. polymyxa proK that contained, respectively, nifB, nifD, and nifK without P2A were used as positive controls. Nitrogenase activity analyses showed that the residual P2A tail had no effect on the activity of NifB and NifD, while P2A tail made NifK to lose activity (Supplementary Figure S2B). The result is consistent with the reports that ENLYFQ tail on K. oxytoca NifK reduced nitrogenase activity in E. coli (Yang et al., 2018). The way to solve this problem is to put NifK on the C-terminal of fusion protein.
Total protein extracts were prepared from the recombinant S. cerevisiae Nif to test expression of nine Nif proteins from two large ORFs. Western blotting with the six antibodies against the proteins NifH, NifD, NifK, NifB, NifE, and NifN of P. polymyxa WLY78 showed that all six Nif proteins were detectable in the protein extracts and had similar sizes with those of P. polymyxa WLY78 ( Figure 2B). The data suggest that Nif proteins could be correctly co-expressed and cleaved in S. cerevisiae. The bands of the proteins (NifD, NifK, NifB, NifE and NifN) are stronger that of NifH. We deduce that the band strength of NifH appear weaker than other proteins (eg. NifB) may be result from its low antibody titer. Since protein numbers are too high, it is hard to estimate the cleavage efficiency mediated by the 2A peptides. Co-expression of four genes was applied to investigate reprogramming by using three 2A peptides ((P2A, T2A, and E2A) to allow expression of the four reprogramming factors in a single vector (Carey et al., 2009). The protein expression level from the first to fourth position in a quad-cistronic construct was characterized and the protein expression gradually decreased toward the 3′ end in quadcistronic constructs (Liu et al., 2017). As we know, 5 proteins linked via 2A in this study are the highest number to be co-expressed in yeast. Finally, we measured the nitrogenase activity of the recombinant S. cerevisiae Nif by using acetylene reduction assays (Wang et al., 2013), but the strain did not show activity. As we know, nitrogenase requires both active NifH and NifDK. Since NifS and NifU for synthesis of Fe-S cluster are required for the active NifH. Thus, we will co-express NifU, NifS and NifH in yeast to check the function of NifH in the follow studies.
To investigate whether NifH was expressed in yeast, total protein extracts were prepared from yeast and analyzed by Western blotting with anti-NifH. The cleaved form NifH (32 kDa) and the uncleaved forms NifS-2A-NifH (75 kDa) and NifU-2A-NifS-2A-NifH (105 kDa) were detectable in recombinant S. cerevisiae USHp and S. cerevisiae USHt, while they were not detected in the recipient S. cerevisiae YSG50 ( Figure 3B). The cleaved NifH (32 kDa) was the major form in yeast and it had the similar molecular size with that of P. polymyxa NifH.
As shown in Figure 3C, the cleavage efficiency of the P2A in S. cerevisiae USHp and in S. cerevisiae USHt was similar at 49-52%, which is lower than the 80% cleavage efficiency in the NifB-P2A-NifH where NifB and NifH were linked by P2A. Our data are consistent with the reports that the highest protein expression was found at the first position in tri-cistronic constructs (Liu et al., 2017). It is also  Saccharomyces cerevisiae USHp cells were initially grown in medium containing 0.6% glucose with strong aeration for about 15 h. After glucose was nearly consumed, 2% galactose was added to culture to induce expression of nif genes and also nitrogen gas was used to induce anerobic fermentation for more than 20 h ( Figure 4A). Western blotting analysis showed that NifH was produced at 21, 25, 27, 29, and 40 h during the fermentation process ( Figure 4B). The cleaved NifH was the major form, and the un-cleaved NifH forms also existed.
His-tagged NifH was purified from S. cerevisiae USHp cells after 40 h of cultivation by using anaerobic Ni 2+ affinity chromatography inside a glove box ( Figure 4C). The purified NifH exhibited brown color which was the characteristic of nitrogenase Fe protein ( Figure 4D). The Fe protein activity of the purified yeast NifH was analyzed by assaying nitrogenase activity in the mixture of yeast NifH and pure K. oxytoca MoFe protein. In absence of yeast NifH, pure K. oxytoca MoFe protein did not exhibit nitrogenase. As observed in positive control of mixtures of pure K. oxytoca Fe and FeMo proteins, yeast NifH showed nitrogenase activity in presence of K. oxytoca MoFe protein. The data indicate that the expressed P. polymyxa NifH in yeast was capable of donating electrons to MoFe protein from K. oxytoca ( Figure 4E). The expressed P. polymyxa NifH in yeast did not require NifM to render active, consistent with the finding that there is no nifM gene in the genome of P. polymyxa. In contrast, A. vinelandii NifH requires NifM to render active Fe protein (López-Torrejón et al., 2016). However, the activity of the purified yeast NifH is 5-10% of that of K. oxytoca Fe protein. We deduce that the low activity of the purified yeast NifH results from two reasons. One reason is that K. oxytoca MoFe protein does not match P. polymyxa Fe protein (NifH) very well. The other reason is that the nitrogenase activity in the mixture of K. oxytoca Fe and MoFe protein is much Frontiers in Microbiology 07 frontiersin.org higher than that in the mixture of P. polymyxa Fe and MoFe. Recently, we have revealed that the nitrogenase activity in the mixture of Fe and MoFe purified from Paenibacillus sabinae T27 is much higher than in the mixture of the purified P. sabinae MoFe and K. oxytoca Fe or A. vinelandii Fe. Also, the nitrogenase activity in the mixture of K. oxytoca Fe and MoFe protein is much higher than that in the mixture of P. sabinae Fe and MoFe (Li et al., 2021a). Recently, we have revealed that the sufCDSUB operon, nifS-like and yutI genes were involved in the Fe-S cluster biosynthesis of nitrogenase in P. polymyxa WLY78 (Li et al., 2021b). In the future, we can co-express nifH and these genes involved in Fe-S cluster biosynthesis of P. polymyxa WLY78 in yeast to investigate whether the Fe activity can be enhanced. In summary, we use 2A peptides to co-express multiple proteins involved in synthesis of the complex nitrogenase system. Our results for the first time demonstrate that Nif proteins linked via 2A peptides can be successfully co-expressed in S. cerevisiae. All the four 2A peptides (P2A, T2A, E2A and F2A) evaluated show cleavage with different efficiency, ranging from ~50% to ~90%. The presence of a 2A tail in NifH, NifB and NifD dose not affect the activity of these Nif proteins, but 2A tail causes a complete loss of NifK activity. Nine Nif proteins (NifH, NifB, NifD, NifK, NifE, NifN, NifX, HesA and NifV) from P. polymyxa WLY78 are successfully co-expressed in yeast. Yeast NifH co-expressed from P. polymyxa NifH and K. oxytoca NifS and NifU exhibits activity of Fe protein.

Strains and media
Bacterial strains and yeast strains used in this study are shown in Supplementary Table S1. S. cerevisiae YSG50 was used as host for nif gene expression. S. cerevisiae Y2H Gold was used as a host in Yeast two hybrid system for detecting the cleavage efficiency of 2A linkers in vivo. S. cerevisiae YSG50 and Y2H Gold strains were usually grown at 30°C in YPD medium (20 g/l tryptone, 10 g/l yeast extract, 20 g/l glucose). The yeast transformants were screened at 30°C in SC medium without tryptophan (SC-Trp), or without tryptophan, histidine and adenine (SC-Trp-His-Ade). E. coli DH5α was used to routine cloning and plasmid propagation. Luria-Bertani broth. P. polymyxa WLY78 was grown in nitrogen-limited medium (Wang et al., 2013). When appropriate, antibiotics were added at the following concentrations: 100 μg/mL Ampicillin and 5 μg/mL Erythromycin for maintenance of plasmids in bacterial strains. Primers and plasmids used in this study were listed in Supplementary Tables S3, S4, respectively.

Construction of Paenibacillus polymyxa nif mutants and other recombinant plasmids
Plasmids for assaying cleavage efficiency of of NifB and NifH fusion protein linked via different 2A peptides in yeast and for yeast two hybrid (Y2H) assay were shown in Supplementary Material. Construction of ΔnifB, ΔnifH, ΔnifD, and ΔnifK mutants of P. polymyxa WLY78 and their complement strains were shown in Supplementary Material.

Yeast two-hybrid assay
Saccharomyces cerevisiae Y2H Gold was a host used in yeast two hybrid. The derivatives of S. cerevisiae Y2H Gold carrying the Nif polyprotein linked via 2A peptide were grown on SC medium lacking tryptophan (SC-Trp) and on SC-Trp-His-Ade agar plates.

Tailing-tolerance assay
To determine the tolerance of Nif protein to the C-terminal tail, nif gene carrying 2A peptide sequence (tail) was used to complement the nif deletion mutant of P. polymyxa WLY78, generating complement strains. Nitrogenase activities of wild-type, mutants and complement strains were comparatively determined.

Western blotting
For immunoblot detection of Nif protein expression and 2A selfcleavage efficiency, S. cerevisiae YSG50 (wild-type) and derivatives were grown in 50 ml flasks containing 20 ml of SD medium until glucose was consumed. Then, 2% galactose was added to induce nif genes separated by 2A sequences expression. Cultures were collected immediately after 9 h of incubation with shaking at 30°C and 200 rpm. Cells of P. polymyxa WLY78 cultivated anaerobically in nitrogen-limited medium were collected and then used as a positive control. The cells pellets were suspended in 200 μL 1 × SDS loading buffer and boiled for 10 min, and then, 20 μL of each sample were loaded onto 12% separating gels. Subsequently, proteins on the gels were transferred to NC membranes (Huaxingbio, China). The antibodies (anti-NifH, anti-NifD, anti-NifK, anti-NifB, anti-NifE, and anti-NifN) were used at a dilution of 1: 10000. The secondary antibody goat anti-rabbit IgG-HRP (Huaxingbio, China) was also used at 1: 10000. Immunochemiluminescene was done by using the NcmECL Western Blot Kit (NCM Biotech, China). The band intensity on the Western blots was analyzed using ImageJ. The cleavage efficiency of 2A linkers was calculated as (cleaved form) / (cleaved form + uncleaved form) × 100%.

Growth of yeast strains and purification of yeast NifH
Saccharomyces cerevisiae USHp was used for purification of NifH protein. It was cultivated in 500 ml of SD medium with shaking at 200 rpm at 30°C as described by Liu et al. (2019). The cells were lysed in a high pressure homogenizer (PhD Technology International LLC, United States) at 25,000 lb. per square inch. His-tagged NifH was purified by Ni 2+ affinity chromatography (QIAGEN, Germany) under anaerobic conditions inside a glovebox (M. Braun Inertgas systems, Germany). Purified NifH protein was stored in liquid N 2 .

Determination of Fe protein activity
Fe protein activity of the yeast NifH was analyzed by assaying nitrogenase activity in vitro. Yeast NifH protein was mixed with pure K. oxytoca MoFe protein, and ATP-regeneration mixture 5 mM ATP, 40 mM creatine phosphate, 10 mM MgCl 2 , 20 mM sodium dithionite, 0.125 mg/mL creatine phosphokinase, 40 mM MOPS-KOH (pH 7.4) (Li and Burris, 1983). Positive control reactions were carried out with pure Fe protein and MoFe protein from K. oxytoca.

Nitrogenase activity assay
Nitrogenase activities of P. polymyxa WLY78 and its derivatives were determined by acetylene reduction assay according to method described by Wang et al. (2018).

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions
SC designed research. MW carried out gene clone and assembly, protein purification, and biochemical studies. XL contributed to gene clone. YS contributed to fermentations. MW and SC analyzed data