A supernumerary synthetic chromosome in Komagataella phaffii as a repository for extraneous genetic material

Abstract

Background: Komagataella phaffii (Pichia pastoris) is a methylotrophic commercially important non-conventional species of yeast that grows in a fermentor to exceptionally high densities on simple media and secretes recombinant proteins efficiently. Genetic engineering strategies are being explored in this organism to facilitate cost-effective biomanufacturing. Small, stable artificial chromosomes in K. phaffii could offer unique advantages by accommodating multiple integrations of extraneous genes and their promoters without accumulating perturbations of native chromosomes or exhausting the availability of selection markers.

Results: Here, we describe a linear "nano"chromosome (of 15-25 kb) that, according to whole-genome sequencing, persists in K. phaffii over many generations with a copy number per cell of one, provided non-homologous end joining is compromised (by KU70-knockout). The nanochromosome includes a copy of the centromere from K. phaffii chromosome 3, a K. phaffii-derived autonomously replicating sequence on either side of the centromere, and a pair of K. phaffii-like telomeres. It contains, within its q arm, a landing zone in which genes of interest alternate with long (approx. 1-kb) non-coding DNA chosen to facilitate homologous recombination and serve as spacers. The landing zone can be extended along the nanochromosome, in an inch-worming mode of sequential gene integrations, accompanied by recycling of just two antibiotic-resistance markers. The nanochromosome was used to express PDI, a gene encoding protein disulfide isomerase. Co-expression with PDI allowed the production, from a genomically integrated gene, of secreted murine complement factor H, a plasma protein containing 40 disulfide bonds. As further proof-of-principle, we co-expressed, from a nanochromosome, both PDI and a gene for GFP-tagged human complement factor H under the control of P_AOX1 and demonstrated that the secreted protein was active as a regulator of the complement system.

Conclusions: We have added K. phaffii to the list of organisms that can produce human proteins from genes carried on a stable, linear, artificial chromosome. We envisage using nanochromosomes as repositories for numerous extraneous genes, allowing intensive engineering of K. phaffii without compromising its genome or weakening the resulting strain.

Keywords: Artificial chromosome; Biomanufacture; Complement factor H; DNA assembly; Metabolic engineering; Pichia pastoris; Therapeutic proteins; Yeast.

Fig. 1

Nanochromosome construction and engineering (also see Additional file 1: Fig. S1). (a) In vitro and in E. coli, we assembled a centromeric, K. phaffii ARS-containing, framework. We added an initial gene-integration array (landing zone) and proto-telomeres, creating nanochromosome (nChr)-precursor version (v)1. We cleaved this between its proto-telomeres, and used the linear product - nChr 1 - to transform K. phaffii cells. To improve nChr stability, we extended nChr-precursor v1 and added a second ARS (for precursor v2), then replaced the initial integration array with new ones, creating precursors v2A and v2B. These yielded nChr 2A and nChr 2B, which were stable only in a ΔKU70 strain. Finally we explored the feasibility of engineering nChr 2A and nChr 2B (to Chr 2A.1 etc.) in vivo. Plasmids are not drawn to scale. (b) Schematic representations (drawn approximately to scale) of the synthetic, linear nanochromosomes as constructed in the current study. Lists of oligos, plasmids and strains may be found in Additional file 2: Tables 1, 2 and 3 respectively. Promoters and terminators are not shown. CEN = centromere, ncDNA = non-coding DNA, LHR = long HR-compatible region, mCH = mCherry gene, GFP:FH = gene coding for a GFP-FH fusion

Fig. 2

Preparation of key parts.(a) Preparation of CEN3, using K. phaffii from gDNA. Left-hand gel: PCR amplicons of segment (Seg) 1 (oligos 197/198), and Seg 2 (oligos 199/200). This approach introduces one extra, functionally silent, base pair (G:C) in the core region (see Additional file 1: Fig. S2). DNA bands of expected sizes (boxed) were extracted and digested with appropriate restriction enzymes before sequential cloning into pUC19 to yield eDA24. Right-hand gel: restriction digestion of eDA24 confirming presence of 6.3-kb CEN3 (IR = inverted repeats). (b) Assembly of framework plasmid in a pUC19-derived scaffold (ori = bacterial origin of replication) (eDA53 validation shown in Additional file 1: Fig. S3). (c) The oligos used to construct the telomeres part (Tel, Additional file 1: Fig. S4). Each proto-telomere contains 16 copies of TGGATGC and the two are linked by the 18-bp I-SceI-recognition site. Lower schematic: cloning of Tel into eDA40 (eDA131 validation shown in Additional file 1: Fig. S5)

Fig. 3

Production and testing of precursor plasmid v1 and nChr 1. (a) Incorporation of an initial, one-off, gene-landing pad into the precursor plasmid. Work on eDA83 was halted due to leaky I-SceI expression in E. coli. (b) Following deletion of I-SceI in eDA83 by insertion of Kan^R (creating eDA110), Tel was inserted to yield eDA137, the precursor plasmid (v1) of nChr 1. The plasmid was linearized in vitro and used to transform (“L&T” in figure) CBS7435 K. phaffii cells creating strain yDA122. The agarose gel shows I-SceI digestion of Tel-containing eDA137 (versus control, i.e. no-Tel, eDA110). (c)  Despite nChr 1 appearing stable in chromosome-loss assays performed on yDA122, WGS (and subsequently, colony PCR, see gel) revealed major chromosomal rearrangements as indicated in the schematic. Note, in yDA122, the loss of PCR-product 3 (of nChr 1), but retention of PCR-product 2 of the suspected translocation product nChr 1* (i.e. nChr 1(p):Chr 3(q))

Fig. 4

Observed aberrations of an extended neochromosome (nChr 2) in wild-type cells (a) The shorter (p) arm of nChr 1 was extended with non-coding (nc)DNA and a second ARS, added by cloning components into precursor v1 to create v2. (b) Agarose gel showing I-SceI digestion products for eDA146 (control) and eDA155. (c) Validation of nChr 2 by PCR before K. phaffii transformation (for oligonucleotides sets see Additional file 2: Table 1). (d) De novo assembly analysis of the WGS for the resultant strains (yDA174 and yDA175) revealed a preponderance of fused chromosomes indicating instability (see Fig. 5 and Additional file 1: Figs. S7 and S8). Strain yDA174 carries a triple-fusion between a nChr 2 q arm (i.e. lacking its p arm), another nChr 2 q arm; and a copy of nChr 2 lacking the p arm telomere; the junction formed between q arms involves the inverted-repeat regions (IRs, 99% sequence identity) of the centromeres but it was not possible to ascertain the nature of this fusion event. In strain yDA175 an additional copy of nChr 2, lacking its p arm telomere, appears fused to the triple-fusion structure seen in yDA174.

Fig. 5

Whole-genome sequencing coverage data for examples of wild-type and ΔKU70 K. phaffii cells containing nanochromosomes (Also see Additional file 1: Fig. S7) The y-axes show “normalised coverage” values obtained by dividing the coverage of each base pair by the mean coverage of all base pairs in the respective sample. This allows comparisons, and identification of duplicated segments or deletions. The x-axes show the position (genomic coordinate) of each base pair in its respective chromosome (values in kilobases for native chromosomes). Each nanochromosome is shown as a schematic (colour-coding as in Fig. 1b), drawn to the same scale as the x-axis in each case. * Indicates a centromere. (.) Indicates anomalies attributable to native genome rearrangements (e.g. duplicated loci) or artifacts arising from highly repetitive regions such as found in telomeres. (X) Indicates a ~ 1-kb duplicated region of Chr 2 corresponding to an ORF encoding an unknown protein (inexplicably observed exclusively in ΔKU70 strains transformed with insertion arrays). Red boxes, drawn on the nanochromosome schematics, indicate CEN3 core regions characterised by the signature presence of an extra G:C base pair. Upper panel: For yDA175, on wild-type background, a plot of normalized coverage indicates two-to-four copies of nanochromosomal sequence per cell. This elevated normalised coverage values for nChr 2 suggest multiple copies per cell of its DNA content. The high copy-number for the Chr 3 centromere correlates with a high copy-number of nChr 2 and supports the existence of chimeric multi-centric nanochromosomes (a tri-centric model is suggested in Additional file 1: Fig. S8). This observation is compatible with yDA175 and yDA177 de novo assembly results obtained from long-read WGS (Additional file 3). Middle and lower panels: Both these strains on a ΔKU70 background are consistent with a single copy per cell of its nanochromosome, specifically: yDA253 with a single copy of nChr 2A; and yDA275 with a single copy of nChr 2A.2 (i.e. after the inch-worming proof-of-principle experiment). [(BioProject PRJNA971544)]

Fig. 6

New versions of precursor plasmids created by swapping out the original integration array. PCR products (see agarose gel) from eDA197 or eDA227, digested with AfeI and XhoI were ligated into appropriately digested eDA155. Precursors v2A and v2B were linearized with I-SceI then used to transform wild-type or ΔKU70 cells to create a total of four strains (see also Additional file 1: Fig. S1)

Fig. 7

Gene expression from the nanochromosome supports production of a traditionally integrated extraneous gene, or can be used directly to produce recombinant protein.(a) Expression of PDI^H on nChr 2A could assist formation of the 40 disulfide bonds of mFH encoded by DNA integrated into Chr 4. (b) SDS-PAGE (Coomassie staining) of crude cell-culture media (30 µL, reducing conditions, treated with Endo H_f) before (d0), and three days after (d3), methanol induction. As expected, co-expression of PDI^H located on either nChr 2A or (as a control) a native chromosome, is required for detectable quantities of mFH to be secreted. (c) A Western blot was used to demonstrate PDI^H production in cells containing nChr 2A three days post-induction. (d) A 165-kDa fusion GFP:FH fusion protein is detectable three days after induction in the cell-culture media from two strains containing nChr 2A.2 carrying both PDI^H and GFP:FH. In each lane 40 µL of EndoH_f–treated crude culture medium (after cell removal) was loaded under reducing conditions. Bands were detected with Coomassie blue and by Western blot using an anti-GFP primary antibody that also revealed the presence of a GFP-containing proteolytic fragment. (e) Both mFH and GFP:FH produced herein are cofactors for (30 nM) complement FI-catalysed cleavage of the 110-kDa α’-chain of complement protein C3b into 63-kD and 39-kDa fragments. In each case 12 µL of crude culture medium (at d(ay) 3 or, as a control, at d(ay) 0) was assayed. A potential contribution of non-specific protease activity is excluded by performing control reactions lacking FI.