Re-evaluation of the impact of BUD21 deletion on xylose utilization by Saccharomyces cerevisiae

Various rational metabolic engineering and random approaches have been applied to introduce and improve xylose utilization and ethanol productivity by Saccharomyces cerevisiae. Among them, the BUD21 gene was identified as an interesting candidate for enhancing xylose consumption as its deletion appeared to be sufficient to improve growth, substrate utilization and ethanol productivity on xylose, even in a laboratory strain lacking a heterologous xylose pathway. The present study aimed at studying the influence of BUD21 deletion in recombinant strains carrying heterologous oxido-reductive xylose utilization pathway. The positive effect of BUD21 gene deletion on aerobic growth and xylose utilization could not be confirmed in two non-engineered laboratory strains (BY4741 and CEN.PK 113-7D) that were grown in YP rich medium with 20 g/L xylose as sole carbon source, despite the fact that effective deletion of BUD21 gene was confirmed using both genotypic (colony PCR) and phenotypic (heat sensitive phenotype of the BUD21 deletion mutant) control experiments. Therefore, the effect of BUD21 deletion on xylose fermentation might be strain- or medium-dependent.


Introduction
Sustainable ethanol production from lignocellulosic feedstock requires complete utilization of all the sugars that are composing cellulose and hemicellulose (Hahn-Hägerdal et al., 2006). This includes xylose that contributes to a substantial carbon portion in some lignocellulosic feedstocks, such as corn stover, hardwood or sugar cane bagasse (Hahn-Hägerdal et al., 2006). Therefore, Saccharomyces cerevisiae, that is the preferred organism for industrial bio-ethanol production, owing to its robustness towards industrially relevant fermentation conditions and ethanol titer (Rudolf et al., 2009), has been engineered for xylose utilization. Rational metabolic engineering and random approaches have been applied to introduce and improve xylose utilization and ethanol productivity by S. cerevisiae (recently reviewed by (Sun and Jin, 2021)). However, xylose utilization is still slow, as compared to glucose, so the search for targets leading to improvement of xylose fermentation rate continues. Several genes in endogenous S. cerevisiae metabolism have been shown to partially or completely suppress the utilization of xylose and hence were considered as deletion targets (Kim et al., 2013a). Some of the reported deletion targets which improved xylose metabolism include YLR042c (non-essential gene encoding a protein with unknown function localized in cell wall) (Parachin et al., 2010;Bengtsson et al., 2008), MNI1 (encoding a putative S-adenosylmethionine-dependent methyltransferase), RPA49 (encoding the α-subunit of RNA polymerase A) (Bengtsson et al., 2008), ALD6 (encoding acetaldehyde dehydrogenase involved in acetate accumulation) (Kim et al., 2013b), PHO13 (encoding an alkaline phosphatase involved in dephosphorylation of xylulose-5-phosphate) (Kim et al., 2013b;Ni et al., 2007), IRA2 and ISU1 (encoding Ras GTPase activating proteins involved in the cAMP/PKA pathway regulation and a mitochondrial Fe-S cluster scaffold protein, respectively) (Sato et al., 2016;Osiro et al., 2019) and BUD21 (encoding a non-essential gene involved in ribosome biogenesis) (Usher et al., 2011). BUD21 (YOR078w) encodes a non-essential nucleolar component associated with U3 snoRNA (that is involved in pre-rRNA processing) required for ribosome biogenesis of small ribosomal subunit processome (SSU) (Shah et al., 2011;Dragon et al., 2002). BUD21 was of particular interest as its deletion in wild type strains S288C and CEN.PK 113-13D was reported to enable xylose utilization, aerobic growth and ethanol production, even in the absence of introduced heterologous xylose utilization pathways (Usher et al., 2011). When a xylose utilization pathway (XI, XKS) was introduced in the BUD21 deletion strain, aerobic fermentation with xylose (20 g/L) was reported to lead to increased amount of xylitol and reduced xylose consumption, which was attributed to a lower XI activity and/or increased expression of GRE3 (Usher et al., 2011), GRE3 being an endogenous aldose reductase gene capable of NADPH-dependent xylose reduction (Träff et al., 2001).
The aim of our study was to investigate the effect of BUD21 gene deletion on xylose utilization in recombinant S. cerevisiae strains carrying, instead, a heterologous oxido-reductive xylose utilization pathway consisting of xylose reductase (XR), xylitol dehydrogenase (XDH) and further optimized by over-expression of gene encoding xylulokinase and several enzymes from the non-oxidative pentose phosphate pathway and deletion of the endogenous GRE3 gene (Johansson and Hahn-Hägerdal, 2002). This could be a strategy to reduce xylitol yield thereby improving ethanol productivity when transformed to industrial S. cerevisiae strains. However, we show here that the initial evaluation of BUD21 deletion in background strains led us to reconsider its importance for xylose utilization.

Generation of deletion strains
Plasmids and strains used in this study are listed in Table 1. BUD21 was deleted in the laboratory strain background CEN.PK 113-7D (Entian and Kötter, 2007) using one-step gene disruption strategy by a fragment generated by overlap-extension PCR. The gene encoding KanMX4 that confers resistance to the antibiotic geneticin (G418) was chosen as selective marker (Güldener et al., 1996). The upstream (US) and downstream (DS) of BUD21 gene were PCR amplified individually with primers BUD21_US_f, BUD21_US_r and BUD21_DS_f, BUD21_DS_r respectively (Table 2) and the PCR products were annealed to the flanking ends of the KanMX4 gene from plasmid pUG6 (Güldener et al., 1996) by overlap-extension PCR utilizing Phusion Hot Start II High-Fidelity DNA Polymerase (Thermo Scientific, Waltham, MA, USA). The resulting fragment (BUD21-US-loxP-KanMX-loxP-BUD21-DS) was purified by QIAquick gel extraction kit and used for transformation. The fragment was transformed into CEN.PK 113-7D strain using the high efficiency LiAc method (Gietz and Schiestl, 2007) and the transformants were selected in YPD plates (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose and 20 g/L agar) supplemented with 200 mg/L geneticin (G418 sulfate) (Gibco, Thermo Scientific, Waltham, MA, USA). BUD21 deletion was confirmed by PCR for at least 3 colonies and with 3 different sets of primers: 1) BUD21_US_f and KanMX_mid_r, 2) KanMX_mid_f and BUD21_DS_r and 3) BUD21_US_f and BUD21_DS_r ( Table 2). Deletion of BUD21 gene in ΔBY4741 strain background was verified using PCR with 3 different sets of primers: 1) BUD21_US_600bp_f and KanMX_mid_r, 2) KanMX_mid_f and BUD21_DS_600bp_r and 3) BUD21_US_f and BUD21_DS_r ( Table 2). The BUD21-deleted strains were designated as ΔBY and ΔCEN.PK (Table 1).

Aerobic growth assays
Growth assay comparisons was carried out between the following S. cerevisiae strains 1) BY (BY4741) and ΔBY (BY4741 strain with BUD21 gene deleted (Accession number: COMP-SET3-A from EUROSCARF deletion mutant library), 2) CEN.PK (wild type CEN.PK strain 113-7D), ΔCEN.PK (CEN.PK 113-7D with BUD21 gene deleted) and TMB3001, a genetically modified xylose-utilizing strain derived from CEN.PK113-7A (Table 1). Pre-inoculum was cultured from single colony of respective strain in 5 mL of YPG medium in 50 mL conical tube incubated for 15-20 h. Based on subsequent cultivation conditions (YPG/YPX media), the CEN.PK, ΔCEN.PK and TMB 3001 strains had an intermediate step where cells were transferred into an inoculum culture of 10 mL YPG/ YPX medium with an initial OD of 0.2 and were incubated for 24 h. The cells from pre-cultures (BY and ΔBY strains) or intermediate cultures (CEN.PK, ΔCEN.PK and TMB3001 strains) were centrifuged at 2050×g for 5 min at 20 • C, washed with 5 mL of 0.9% sterile NaCl solution and the harvested cells were re-inoculated into 25 mL of YPG/YPX media or YNG/YNX media (for BY and ΔBY strains) in 250 mL baffled Erlenmeyer shake flasks with an initial OD of 0.2. Aerobic growth was performed at 30 • C in a rotary shake incubator (New Brunswick, Enfield, CT, USA) at 180 rpm. Absorbance measurements were performed at 620 nm to determine cell growth on a given sample diluted to an optical density (OD) below 0.3 (Spectrophotometer U-1800, Hitachi, Berkshire, UK).

Heat sensitivity assay
In order to confirm the phenotypic difference between CEN.PK and ΔCEN.PK, a growth curve was performed in YPG medium at 37 • C. Preinoculum was cultured from single colony of respective strains in 5 mL of YPG medium in 50 mL conical tube incubated for 18 h at 30 • C. The cells from these pre-cultures were re-inoculated into 25 mL of YPG medium in 250 mL baffled Erlenmeyer shake flasks with an initial OD of 0.2 and grown at 37 • C and 180 rpm.
All the aerobic cultivations were performed at least in biological duplicates. Student T-test was performed with two tailed distribution and two sample unequal variance using Microsoft excel 2013 software.

Metabolite analysis
Samples were withdrawn for metabolite analysis, quickly

Results and discussion
Before investigating the effect of BUD21 gene deletion on xylose utilization in strains carrying the Scheffersomyces stipitis oxidoreductive xylose pathway, a preliminary experiment was performed to confirm the previously observed response of a non-engineered laboratory strains to BUD21 gene deletion, i.e. the ability to grow on xylose without added heterogeneous xylose genes (Usher et al., 2011). Aerobic growth on xylose of the BUD21 deletion mutant from EUROSCARF collection and the parental strain BY4741 (Table 1) were first compared in YPX medium. As BUD21 deletion was also reported to improve growth on glucose (Usher et al., 2011), the comparison was also performed in YPG medium. In parallel, proper deletion of BUD21 gene in the ΔBY strain background was confirmed using PCR (data not shown), which ruled out potential strain construction issues.
Deletion of BUD21 led to a marginal improvement in the maximum specific growth rate on xylose (Table 3, Fig. 1) as compared with wild type BY strain (P-value = 0.02); however, the effect was not as significant as previously reported (Usher et al., 2011) and the final OD values were similar (Table 3). When grown in glucose medium (YPG), ΔBY strain had a similar final OD compared to the wild type BY strain (P-value = 0.19) but a lower maximum specific growth rate than in wild type strain (P-value = 0.0001) ( Table 3).
Yeast extract that is used in culture media may contain traces of glucose obtained, for instance, from glycogen and trehalose hydrolysis as well as amino acids and other biosynthetic precursors that are essential for anabolic pathways, thereby saving metabolic energy which can translate into improvement in growth (Hahn-Hägerdal et al., 2005). As the growth rate and final OD were poor but significant in YPX, we investigated the impact of the absence of yeast extract on the strain performances. BY and ΔBY strains were grown aerobically in defined mineral medium with glucose (YNG) or xylose (YNX) as unique carbon sources. As expected, the growth rate was lower in YNG than in the YPG medium for both strains. But, most important, similar growth profiles were obtained for BY and ΔBY strains in YNG medium, whereas no growth was observed for both strains in YNX medium (Fig. 2). This indicates that the marginal growth of BY and ΔBY strains in YPX medium   Fig. 1. Representative aerobic growth of S. cerevisiae strains BY (squares) and ΔBY (triangles) in YPG (filled symbols) and YPX media (empty symbols). Experiments were performed in duplicates and less than 5% deviation was recorded.

Fig. 2.
Representative aerobic of S. cerevisiae strains BY (squares) and ΔBY (triangles) in YNG (filled symbols) and YNX media (empty symbols). Experiments were performed in duplicates and with less than 5% deviation was recorded.
might have not resulted from the use of xylose as carbon source and it supports previous work where growth of S. cerevisiae was observed in yeast extract medium lacking any added carbon source (Hahn-Hägerdal et al., 2005). Since this first result was unexpected, it was decided to delete BUD21 in another strain background in order to evaluate whether BUD21associated improvement could be strain specific. CEN.PK 113-7D (CEN. PK) was chosen to re-analyze the effect of BUD21 deletion on xylose metabolism since it is a popular strain for physiological studies (van Dijken et al., 2000) and it is the same background strain as for the previous BUD21 deletion study (Usher et al., 2011). A BUD21-deletion fragment was generated by overlap-extension PCR, and transformed into CEN.PK 113-7D and the successful deletion was confirmed using colony PCR towards the gene locus (data not shown). The BUD21 deleted strain (ΔCEN.PK) was grown aerobically in YPX medium with the parental strain CEN.PK as control strain and YPG as control medium. The fermentation metabolite data indicated that xylose consumption was below detection limits (data not shown) in both CEN.PK and ΔCEN.PK strains, which demonstrated the absence of xylose utilization by both strains. This was also confirmed by the limited and identical growth profile in both strains (Fig. 3), that is expected to arise from the YP component of the medium, as discussed above. In order to validate the suitability of the used medium for growth on xylose, the xylose-utilizing control strain TMB 3001 (Eliasson et al., 2000) was cultivated in both YPG and YPX media. As expected, TMB 3001 grew in both media (Fig. 3) and used glucose in YPG (Fig. 4A) and xylose in YPX (Fig. 4B), with consumption rates of 2.94 ± 0.00 g L − 1 .h − 1 and 0.09 ± 0.01 g L − 1 .h − 1 on YPG and YPX, respectively; In contrast, no xylose consumption was measured in YPX medium for the BUD21 deleted version of CEN.PK 113-7D strain (Fig. 4B).
As a last control to verify the correct inactivation of BUD21 gene, CEN.PK and ΔCEN.PK strains were grown in YPG medium at 37 • C as the deletion of BUD21 has been reported to decrease heat tolerance at elevated temperature (37 • C) (Sinha et al., 2008). Maximum specific growth rate of the ΔCEN.PK strain (0.28 h − 1 ) indeed decreased 1.5 fold as compared with the CEN.PK strain (0.43 h − 1 ) (P-value = 0.022) (Fig. 5), which confirmed the deletion phenotype. Consequently, it was not considered relevant to perform further deletion of BUD21 in recombinant strains carrying the oxido-reductive xylose pathway.
BUD21 is a 645 nucleotide ORF coding for a protein of 214 amino acids involved in ribosomal biogenesis processes, including production of 18S RNA, rRNA processing and 40S ribosome subunit assembly (Dragon et al., 2002;Granneman and Baserga, 2004). As ribosome biosynthesis is an extremely energy demanding process in rapidly growing cells (Shah et al., 2011), optimization or deletion of some non-essential ribosomal biogenesis pathway genes, that would reduce ribosomal protein expression, could be a strategy to grow and ferment S. cerevisiae strains under specific stress conditions that require the mobilisation of cell resources (Shah et al., 2011). This way, more resources and energy could be conserved and channelled to cell growth and tolerance during episodes of stress. It was indeed proved that BUD21 gene mutation led to increased tolerance to hypoxia (Shah et al., 2011) and we could confirm the growth sensitivity at elevated temperature. Nevertheless, we could not corroborate that BUD21 gene was involved in the initiation and improvement of xylose utilization from our aerobic growth studies.
Our results highlight the importance of the choice of the medium for the assessment of genetic manipulations. It was previously shown that baker's yeast can grow aerobically by utilizing yeast extract as the sole Fig. 3. Representative aerobic growth of S. cerevisiae strains CEN.PK (squares), ΔCEN.PK (triangles) and TMB 3001 (circles) in YPG (filled symbols) and YPX media (empty symbols). Experiments were performed in duplicates and less than 5% deviation was recorded. glucose/xylose (diamond), glycerol (squares), acetate (triangle), ethanol (circle) and xylitol (cross). Experiments were performed in duplicates with less than 5% deviation.

Fig. 5.
Representative aerobic growth of S. cerevisiae strains CEN.PK (squares) and ΔCEN.PK (triangles) in YPG at 37 • C. Experiments were performed in duplicates with less than 5% deviation. carbon source, with a maximum specific growth rate of 0.29 h − 1 and final OD of 3-4 (Hahn-Hägerdal et al., 2005). It is therefore possible that the growth previously observed for BUD21 deletion strains in YPX medium originated from the YP components and not from xylose itself. We have recurrently observed that different YP batches may contain different levels of remaining sugars, which impacts the interpretation of growth data. The study re-inforce the fact that comparison studies should be performed in defined mineral media, such as the one reported by (van Dijken et al., 2000).
Since there were slight difference in growth in YPX medium between strains of BY4741 and CEN.PK 113-7D, it is possible that strain background also plays a role in the growth phenotype. For instance, the improvement of xylose utilization was previously achieved by deleting YLR042c in different laboratory strain backgrounds but the effect was shown to be the most significant in the strain having the poorest growth rate (Parachin et al., 2010;Bengtsson et al., 2008). The same phenomenon might be applicable here with BUD21 and the improvement of growth, which we attribute to the residual carbon source that is present in yeast extract, might be significant in terms of fold change but limited in terms of achieved biomass as well as mostly visible in strains having poor growth capacity.

Conclusions
BUD21 gene deletion was previously reported to enable aerobic growth and xylose utilization in S. cerevisiae (Usher et al., 2011). This phenotype could not be reproduced in two different non-engineered laboratory strain backgrounds (BY4741 and CEN.PK 113-7D) and on different media, despite the fact that effective deletion of BUD21 gene was confirmed. Therefore, the effect of BUD21 deletion on xylose fermentation might be strain-or medium-dependent.