Deletion of Budding Yeast MAD2 Suppresses Clone-To-Clone Differences in Artificial Linear Chromosome Copy Numbers and Gives Rise

microorganisms

Article Deletion of Budding Yeast MAD2 Suppresses Clone-to-Clone Diﬀerences in Artiﬁcial Linear Chromosome Copy Numbers and Gives Rise to Higher Retention Rates

1,2, 1, 1, 1 1 Scott C. Schuyler * , Lin-Ing Wang † , Yi-Shan Ding †, Yi-Chieh Lee and Hsin-Yu Chen 1 Department of Biomedical Sciences, College of Medicine, Chang Gung University, Kwei-Shan, Tao-Yuan 333, Taiwan; [email protected] (L.-I.W.); [email protected] (Y.-S.D.); [email protected] (Y.-C.L.); [email protected] (H.-Y.C.) 2 Division of Head and Neck Surgery, Department of Otolaryngology, Chang Gung Memorial Hospital, Kwei-Shan, Tao-Yuan 333, Taiwan * Correspondence: [email protected]; Tel.: +886-3-211-8800-x3596 These authors contributed equally to this work. † Received: 26 August 2020; Accepted: 25 September 2020; Published: 29 September 2020

Abstract: Our goal was to investigate the changes in artificial short-linear chromosome average copy numbers per cell arising from partial or full loss of Mitotic Arrest-Deficient 2 (MAD2) spindle checkpoint function in budding yeast Saccharomyces cerevisiae. Average artificial linear chromosome copy numbers in a population of cells, as measured by quantitative polymerase chain reactions (qPCR), and retention rates, as measured by fluctuation analyses, were performed on a total of 62 individual wild type and mad2∆ mutant haploid and diploid clones. Wild type cells, both haploids and diploids, displayed phenotypically unique clone-to-clone differences: one group of 15 clones displayed low-copy numbers per cell and high retention rates, were 1 clone was found to have undergone a genomic integration event, and the second group of 15 clones displayed high copy numbers per cell and low retention rates, with the latter values being consistent with the previously published results where only a single clone had been measured. These chromosome states were observed to be unstable when propagated for 10 days under selection, where high copy-low retention rate clones evolved into low copy-high retention rate clones, but no evidence for integration events was observed. By contrast, mad2∆ haploid and mad2∆/mad2∆ diploids displayed a suppression of the clone-to-clone differences, where 20 out of 21 clones had mid-level artificial linear chromosome copy numbers per cell, but maintained elevated chromosome retention rates. The elevated levels in retention rates in mad2∆ and mad2∆/mad2∆ cells were also maintained even in the absence of selection during growth over 3 days. MAD2/mad2∆ heterozygous diploids displayed multiple clonal groups: 4 with low copy numbers, 5 with mid-level copy numbers, and 1 with a high copy number of artificial linear chromosomes, but all 10 clones uniformly displayed low retention rates. Our observations reveal that MAD2 function contributes to the ability of yeast cells to maintain a high number of artificial linear chromosomes per cell in some clones, but, counter-intuitively, mad2∆ suppresses clone-to-clone differences and leads to an improvement in artificial linear chromosome retention rates yielding a more uniform and stable clonal population with mid-level chromosome copy numbers per cell.

Keywords: budding yeast; Saccharomyces cerevisiae; chromosome; mitotic spindle checkpoint; Mitotic Arrest-Deﬁcient 2 (MAD2); quantitative polymerase chain reaction (qPCR)

Microorganisms 2020, 8, 1495; doi:10.3390/microorganisms8101495 www.mdpi.com/journal/microorganisms Microorganisms 2020, 8, 1495 2 of 15

1. Introduction There is broad interest in using budding yeast artificial short-linear chromosomes as a tool to study the mechanisms of chromosome segregation in mitosis and as a technology to carry large fragments of exogenous DNA for recombinant protein expression, for bioengineering of entire metabolic pathways, where all of the pathway enzymes are carried on a single large artificial chromosome, or for amplification of exogenous DNA clones, such as fragments of the human or mouse genomes or other organisms [1,2]. Artificial linear chromosomes also have the advantage of containing telomeres, a centromere, and an origin of replication all of which can contribute to a high rate of chromosome segregation fidelity relative to other smaller circular plasmids that lack all of these elements [3–7]. These chromosome elements also allow artificial linear chromosomes to maintain a presence in a population of yeast even in the absence of selection, at least for short durations of time. When working with exogenous DNA elements in microorganisms, such as plasmids, scientists must be mindful of the DNA copy numbers per cell. Both as a tool to study mitosis, and as a technology to carry large fragments of exogenous DNA, the average copy numbers of artificial chromosomes per cell might have an impact on optimal production or yields of desired target biomolecules or biological products such as during large-scale fermentation of yeast cultures. Average linear chromosome numbers in populations of yeast cells have been measured for individual clones in the past using Southern blots with an external piece of plasmid DNA as a reference standard, and retention rates, as measured by classic fluctuation analyses [6,7]. Using these approaches, average copy numbers per yeast cell were calculated as the average chromosome numbers divided by the retention rates, where values in the range of 7.5 artificial linear chromosomes per yeast cell with a retention rate of 88% or 15 artificial linear chromosomes per yeast cell with a retention rate of 61% in haploids were reported [6,7]. One example of a clone with an average of 30 artificial linear chromosomes in a diploid yeast that displayed a retention rate of only 30% has also been reported [7]. These quantitative studies were performed in wild-type yeast cell backgrounds containing all mitotic and cell cycle regulatory pathways which, in principle, should maximize chromosome segregation fidelity, but were limited to single clones [6,7]. To ensure chromosome segregation fidelity, including the segregation of artificial linear chromosomes, eukaryotic cells contain a highly-conserved mitotic spindle checkpoint pathway that promotes accurate chromosome segregation by delaying cell cycle progression until all chromosomes become attached at their kinetochores and are placed under tension between the two sister-kinetochores of paired sister chromatids held together by cohesion rings in opposition to the forces generated by the mitotic spindle before the metaphase-anaphase transition [8–11]. If a mitotic spindle microtubule attaches to a kinetochore, but physical tension is not built up by the spindle forces between the paired sister kinetochores, the microtubule attachment becomes unstable and is lost, leaving an unattached chromosome, which activates the spindle checkpoint. One core component of the checkpoint pathway is the Mitotic Arrest-Deficient 2 (MAD2) gene that ensures chromosome segregation stability so that each daughter cell inherits a single copy of each chromosome [8,9,11]. Mad2 protein is recruited to unattached kinetochores by interacting with the Mitotic Arrest-Deficient 1 (Mad1) protein where it undergoes a dramatic conformational change and becomes bound to Cell Division Cycle 20 (Cdc20) protein [10,11]. This interaction recruits Mitotic Arrest-Deficient 3 (Mad3) and Budding Uninhibited by Benomyl 3 (Bub3) to form the Mitotic Checkpoint Complex (MCC) which blocks the activity of an E3 ubiquitin Ligase, the Anaphase-Promoting Complex/Cyclosome (APC/C) [10,11]. APC/CCdc20 enzyme function is essential to initiate the metaphase-anaphase transition by promoting the ubiquitin-dependent proteolysis of core cell cycle regulators like B-type Cyclins and Securin, an inhibitor of anaphase [10,11]. In budding yeast, the removal of the MAD2 function destroys spindle checkpoint function. Loss in the Mad2 function creates a failure to form the Mitotic Checkpoint Complex, allowing the APC/CCdc20 to remain active, leading to the pre-mature destruction of B-type Cyclins and Securin which allows progression into anaphase even in the absence of chromosome attachments to the mitotic spindle [10,11]. Short-linear artificial chromosomes present the yeast spindle checkpoint pathway with an additional unique challenge: the forces exerted on the short chromosomes has been proposed Microorganisms 2020, 8, 1495 3 of 15 to be able to induce premature separation of sister chromatids at a low frequency before the metaphase-anaphase transition, thus making it impossible for the development of tension and proper stable microtubule-chromosome attachments on the pre-maturely separated artificial linear chromosomes, leading into an increase in chromosome mis-segregation rates [12]. Based on our past work, we chose to try and investigate further the consequences of damaging or destroying mitotic spindle checkpoint function on artificial linear chromosome segregation fidelity [12]. Our goal was to investigate the changes in the average number of artificial linear chromosomes per cell arising in response to the loss of MAD2 spindle checkpoint function in budding yeast. Towards this end, we developed a new quantitative polymerase chain reaction (qPCR) strategy to measure the average artificial linear chromosome copy numbers, and also measured retention rates by classic genetic fluctuation analyses, in haploid and diploid cells where MAD2 was a wild type, partially reduced or deleted. Initial measurements in wild type MAD2 cells yielded an unanticipated clone-to-clone difference in chromosome copy numbers from one genetically identical clone compared to another. To explore this clone-to-clone difference in chromosome copy number phenotype further, we expanded our analyses to include 62 individual clones in total and observed that MAD2 function contributes to the maintenance of a high number of artificial linear chromosomes in some clones, but, counter-intuitively, mad2∆ haploid and mad2∆/mad2∆ diploid clones display improved artificial chromosome retention rates and suppress clone-to-clone differences in chromosome copy numbers to yield a different phenotype of clones with mid-level chromosome copy numbers per cell with high retention rates.

2. Materials and Methods Standard yeast protocols were performed for the growth and transformation of yeast cells as previously described [13]. The strains used in this study are listed in the Supplementary Materials, and strains and plasmids are available upon request (Table S1). Cells were transformed with circular plasmids as a control, or short-linear artificial chromosomes selected for on complete synthetic media minus leucine (CSM-leu). Circular and linear artificial chromosomes carried either the LEU2 or LEU2-∆16 leucine auxotrophic marker for selection on CSM-leu media, where LEU2-∆16 is a wild type leucine auxotrophic allele containing a 16 base pair deletion upstream of the ATG start codon of the LEU2 open-reading frame which has no observable effect on mRNA expression levels of the LEU2 open-reading frame (Figure S1) [7,14,15]. Retention rate assays were performed by growing cells overnight in selective CSM-leu media, and then plating them onto either CSM or CSM-leu plates and counting colonies [6,7]. Briefly, after growth overnight, and OD600 was measured to determine the number of cells per milliliter, and then cells were washed and diluted to plate at an approximate density of 100 cells per plate to allow for colony counting. Cells were grown for 2–3 days at 30 ◦C to allow colonies to grow. Each retention assay measurement for each clone was made 6 times to ensure a low standard deviation and increased confidence in the measured values. Statistical analyses were performed using a GraphPad Prism calculator employing the Student’s unpaired t-test (GraphPad, San Diego, CA, USA). The execution of quantitative polymerase chain reaction (qPCR) chromosome copy number genomic DNA measurements were made by using DNA isolated from cells with TRIzol reagent (ThermoFisher Scientific, Waltham, MA, USA), and qPCR was performed following the manufacturer’s recommends using SYBR GREEN (Roche, San Francisco, CA, USA) and the StepOnePlus Real-Time PCR System (ThermoFisher Scientific, Waltham, MA, USA). Briefly, reaction template DNA and primer concentrations were adjusted, where untransformed wild type haploid (SCSY79) or diploid (SCSY1037) genomic DNA was used as a reference control to obtain Cycle-threshold (Ct) values of about 20 to ensure clear detection of fluorescent SYBR GREEN labeled PCR products as recommended by the manufacturer (ThermoFisher Scientific, Waltham, MA, USA). Multiple combinations of primer pairs were initially tested to determine which pairs gave rise to a single, well-amplified classic PCR product, which is a necessary prerequisite for successful qPCR. Primer pairs selected in this manner for qPCR were “CUP9 forward” (50-GCAAAGCCCTGCTACAAATCC-30) and “CUP9 reverse” Microorganisms 2020, 8, 1495 4 of 15

(50-AGACCTCCTGCCCGAATTTT-30); “ILV6-2 forward” (50-CAGTGCCGTTTCATCCATCA-30) and “ILV6-2 reverse” (50-TCTTTGACCTCGGTGTTGCA-30); “LEU2-4 forward” (50-CATGAGCCA CCATTGCCTATT-3’) and “LEU2-4 reverse” (50-TGGCGGCAGAATCAATCAA-30). Primer pairs were also designed to have identical melting temperatures for qPCR. The “LEU2-4 forward” and “LEU2-4 reverse” primer pair amplifies the wild type LEU2 locus and the endogenous leu2-3,112 chromosomal mutant locus equally well. To establish the linear range for the qPCR reactions, initial qPCR assays were performed using recombinant DNA plasmids instead of yeast genomic DNA, where we observed that the qPCR measurements remained linear under our conditions within a range of 1–15 measured copies relative to a standard amount of DNA set equal to 1. A typical qPCR reaction contained 1–2 µL of genomic template DNA that had been quantified and tested by classic PCR first to make sure it could serve as a good quality high-PCR yield template, 1.6 µL of the forward primer, 1.6 µL of the reverse primer, 10 µL of the SYBR GREEN reaction mix, where the remaining volume was adjusted to 20 µL with ddH2O. If the genomic template DNA was found to be of low quality it was discarded, and new samples were prepared, where each new sample was examined by classic PCR first to ensure its quality as a template yielding only a single well-amplified PCR product. For the initial measurements, qPCR was performed 16 times on genomic DNA samples isolated from each individual haploid clone (8 clones total, meaning 128 individual qPCR reactions) and 12 times on genomic DNA isolated from each individual diploid clone (6 clones total, meaning 72 individual qPCR reactions) to ensure that a calculated final artificial linear chromosome copy number per cell value had a standard deviation of less than 1 (Table S1). However, to measure a larger number of clones (62 artificial linear chromosome clones in total) to further investigate the observed clone-to-clone differences, qPCR reactions were performed only 8 times on the genomic DNA isolated from each individual clone to minimize the use of reagents but still achieve a low standard deviation and high confidence in the measured values for each individual clone sample, where there is in general good agreement between the measurements made using 16, 12, and 8 replicates of the qPCR reactions on the genomic DNA isolated from each clone (Table S2). All data for retention rates, qPCR measurements, and calculated artificial linear chromosome copy numbers are listed for each yeast clone (Table S3).

3. Results

3.1. Short Linear Artificial Chromosomes Display Clone-to-Clone Differences in Wild Type Haploid Cells Our initial goal was to investigate the potential role of the mitotic checkpoint gene MAD2 in the maintenance of artificial linear chromosomes. We employed the classic genetic fluctuation test and developed a qPCR-based method to determine the artificial linear chromosome copy numbers per cell in a population (Figure1). We initially applied the assay to measure the average number of artificial circular plasmids per cell as a control, where previous reports have established circular chromosomes are carried at 1–2 copies per cell on average [3–7]. Our initial results were in good agreement with this, where average chromosome copy numbers ranged between (0.96 0.23) up to (2.19 0.96) in a variety ± ± of wild type MAD2 and mutant mad2∆ haploid and diploid clones (Table S1). Having established our assay, we initially measured the average artificial linear chromosome copy numbers in 11 genetically identical wild type haploid clones and detected an unexpected phenotype among the clones (Figure2). One set of 5 clones displayed a high copy number of artificial linear chromosomes (average of 6.81 1.97) and lower retention rates (average of 68.8 6.6%), while the ± ± second set of clones contained a significantly lower copy number of artificial linear chromosomes (average of 1.91 0.71; p-value < 0.001) and had significantly higher retention rates (average of ± 79.2 7.17%; p-value < 0.05). ± Microorganisms 2020, 8, x FOR PEER REVIEW 4 of 15

“ILV6-2 forward” (5’-CAGTGCCGTTTCATCCATCA-3’) and “ILV6-2 reverse” (5’- TCTTTGACCTCGGTGTTGCA-3’); “LEU2-4 forward” (5’-CATGAGCCACCATTGCCTATT-3’) and “LEU2-4 reverse” (5’-TGGCGGCAGAATCAATCAA-3’). Primer pairs were also designed to have identical melting temperatures for qPCR. The “LEU2-4 forward” and “LEU2-4 reverse” primer pair amplifies the wild type LEU2 locus and the endogenous leu2-3,112 chromosomal mutant locus equally well. To establish the linear range for the qPCR reactions, initial qPCR assays were performed using recombinant DNA plasmids instead of yeast genomic DNA, where we observed that the qPCR measurements remained linear under our conditions within a range of 1–15 measured copies relative to a standard amount of DNA set equal to 1. A typical qPCR reaction contained 1–2 µL of genomic template DNA that had been quantified and tested by classic PCR first to make sure it could serve as a good quality high-PCR yield template, 1.6 µL of the forward primer, 1.6 µL of the reverse primer, 10 µL of the SYBR GREEN reaction mix, where the remaining volume was adjusted to 20 µL with ddH2O. If the genomic template DNA was found to be of low quality it was discarded, and new samples were prepared, where each new sample was examined by classic PCR first to ensure its quality as a template yielding only a single well-amplified PCR product. For the initial measurements, qPCR was performed 16 times on genomic DNA samples isolated from each individual haploid clone (8 clones total, meaning 128 individual qPCR reactions) and 12 times on genomic DNA isolated from each individual diploid clone (6 clones total, meaning 72 individual qPCR reactions) to ensure that a calculated final artificial linear chromosome copy number per cell value had a standard deviation of less than 1 (Table S1). However, to measure a larger number of clones (62 artificial linear chromosome clones in total) to further investigate the observed clone-to- clone differences, qPCR reactions were performed only 8 times on the genomic DNA isolated from each individual clone to minimize the use of reagents but still achieve a low standard deviation and high confidence in the measured values for each individual clone sample, where there is in general good agreement between the measurements made using 16, 12, and 8 replicates of the qPCR reactions on the genomic DNA isolated from each clone (Table S2). All data for retention rates, qPCR measurements, and calculated artificial linear chromosome copy numbers are listed for each yeast clone (Table S3).

3. Results

3.1. Short Linear Artificial Chromosomes Display Clone-to-Clone Differences in Wild Type Haploid Cells Our initial goal was to investigate the potential role of the mitotic checkpoint gene MAD2 in the maintenance of artificial linear chromosomes. We employed the classic genetic fluctuation test and developed a qPCR-based method to determine the artificial linear chromosome copy numbers per cell in a population (Figure 1). We initially applied the assay to measure the average number of artificial circular plasmids per cell as a control, where previous reports have established circular chromosomes are carried at 1–2 copies per cell on average [3–7]. Our initial results were in good Microorganismsagreement 2020with, 8 this,, 1495 where average chromosome copy numbers ranged between (0.96 ± 0.23) up5 of to 15 (2.19 ± 0.96) in a variety of wild type MAD2 and mutant mad2∆ haploid and diploid clones (Table S1).

Microorganisms 2020, 8, x FOR( aPEER) REVIEW (b) 5 of 15

FigureFigure 1.1. MeasurementsMeasurements inin a population of yeast of th thee average average artificial artificial chromosome chromosome copy copy numbers numbers perper cell: cell: (a(a)) Genetic Genetic fluctuation fluctuation assays assays were were performed performed by by initially initially growing growing clones clones in in selective selective CSM-leu CSM- medialeu media overnight overnight and thenand then plating plating them them on CSM on CS andM CSM-leuand CSM-leu to measure to measure the averagethe average retention retention rates (3rates/5 = (3/560% = in60% this in example);this example); (b) ( Ab) schematicA schematic diagram diagram of of the the qPCR qPCR assayassay toto measure the the number number ofof copiescopies of the leu2-3,112leu2-3,112 ++ LEU2LEU2 lociloci present present in the in theisolated isolated genomic genomic DNA DNA from froma population a population of cells. of cells.The IVL6 The locusIVL6 onlocus chromosome on chromosome III and IIICUP9 and locusCUP9 onlocus chromosome on chromosome XVI served XVI as served internal as controls internal controlsthat remain that at remain an equal at an1:1 equalratio with 1:1 ratio each with other each and otherwith the and endogenous with the endogenous leu2-3,112 locus,leu2-3,112 wherelocus, the wheresmall thevertical small lines vertical represent lines represent the qPCR the qPCRproducts products.. PCR PCRprimer primer pairs pairs were were designed designed toto ensure ensure amplificationamplification of of bothbothleu2-2,112 leu2-2,112 andand LEU2LEU2 at the same level. In In this this ex example,ample, the the qPCR qPCR cycle-threshold cycle-threshold (Ct)(Ct) for for the theleu2-3,112 leu2-3,112+ + LEU2 locus primer pair pair would would occur occur earlier earlier relative relative to to an an untransformed untransformed wild wild typetype cell cell populationpopulation containingcontaining onlyonly the single leu2-3,112 locuslocus on on chromosome chromosome III III (the (the schematic schematic is is meantmeant toto representrepresent aa populationpopulation ofof cellularcellular genomic DNA containing containing an an average average of of 5 5 artificial artificial linear linear chromosomeschromosomes –– inin aa realreal populationpopulation mademade up of many individual cells, cells, some some cells cells would would carry carry more more artificialartificial chromosomes chromosomes andand othersothers wouldwould carry fewer artificial artificial chromosomes); chromosomes); ( (cc)) in in all all qPCR qPCR reactions reactions untransformeduntransformed wildwild typetype haploidhaploid or diploid cells we werere employed employed as as reference reference controls controls where where the the ratio ratio betweenbetween thethe leu2-3,112leu2-3,112 andand ILV6ILV6 and CUP9 internal controls controls could could be be set set to to equal equal 1, 1, and and then then were were employedemployed to to normalize normalize the the measured measuredILV6 ILV6and CUP9and CUP9signals signals from qPCR from reactions qPCR reactions in samples in containing samples artificialcontaining linear artificial chromosomes linear chromosomes (alc). The average (alc). The chromosome average chromosome counting number counting is defined number as is thedefined qPCR signalas the measuredqPCR signal relative measured to the relativeILV6 and to theCUP9 ILV6internal and CUP9 controls, internal minus controls, 1 in haploidsminus 1 in to haploids remove theto over-countremove the of over-count the endogenous of the endogenousleu2-3,112 locus; leu2-3,112 (d) Calculating locus; (d) theCalculating average artificialthe average linear artificial chromosome linear copychromosome numbers copy per cell.numbers The qPCR per cell. amplifies The qPCR genomic amplifies DNA genomic isolated DNA from isolated a population from ofa population cells which willof cells include which cells will that include have very cells recently that have lost very the artificialrecently linearlost the chromosome artificial linear (for example,chromosome in the (for last mitoticexample, division), in the last even mitotic when division), under selection, even when but thatunder have selection, not yet but died that (in have this example,not yet died 2 cells (in havethis lostexample, the artificial 2 cells chromosomeshave lost the artificial and 3 cells chromosome have retaineds and them;3 cells wherehave retained a total of them; 20 leu2-3,112 where a total+ LEU2 of loci20 leu2-3,112 would be + amplified LEU2 loci during would qPCRbe amplified off of the during genomic qPCR DNA off of of the these genomic 5 cells; DNA the over-count of these 5 cells; of the leu2-3,112the over-countsignal of would the leu2-3,112 be subtracted signal outwould as describedbe subtracted above out in as ( cdescribed) leaving aabove total in of ( 15c) leavingLEU2 loci a distributedtotal of 15 LEU2 among loci genomic distributed DNA among from 5 genomic cells giving DNA rise from to an 5 cells underestimate giving rise of to an an average underestimate of 3 copies of ofanLEU2 averageper cell).of 3 copies The retention of LEU2 rate per reveals cell). The the trueretention number rate of cellsreveal ins thethe populationtrue number that of have cells retained in the thepopulation artificial chromosomethat have retained at the th momente artificial of isolation chromosome of the at genomic the moment DNA of (3 isolation cells in this of examplethe genomic with DNA (3 cells in this example with 9, 5, or 1 copy(ies) of the artificial linear chromosome). The true average copy number of artificial linear chromosomes in the population of cells that have retained the chromosomes is determined by normalizing the measured qPCR average copy number by the number of genomes counted as quantified by the amount of the genomic DNA added as a template to the qPCR reaction sample (5 in this example) divided by the retention rate (3/5 in this example) to yield the final result: an average of 5 copies of the artificial linear chromosomes per cell.

Having established our assay, we initially measured the average artificial linear chromosome copy numbers in 11 genetically identical wild type haploid clones and detected an unexpected phenotype among the clones (Figure 2). One set of 5 clones displayed a high copy number of artificial linear chromosomes (average of 6.81 ± 1.97) and lower retention rates (average of 68.8 ± 6.6%), while the second set of clones contained a significantly lower copy number of artificial linear chromosomes (average of 1.91 ± 0.71; p-value < 0.001) and had significantly higher retention rates (average of 79.2 ± 7.17%; p-value < 0.05).

Microorganisms 2020, 8, 1495 6 of 15

9, 5, or 1 copy(ies) of the artificial linear chromosome). The true average copy number of artificial linear chromosomes in the population of cells that have retained the chromosomes is determined by normalizing the measured qPCR average copy number by the number of genomes counted as quantified by the amount of the genomic DNA added as a template to the qPCR reaction sample (5 in this example) divided by the retention rate (3/5 in this example) to yield the final result: an average of

5 copiesMicroorganisms of the 2020 artiﬁcial, 8, x FOR linear PEER chromosomesREVIEW per cell. 6 of 15

(a) (b)

FigureFigure 2. Genetically 2. Genetically identical identical wild wild typetype MAD2MAD2 haploidhaploid clones clones have have an unexpected an unexpected phenotype phenotype displayingdisplaying clone-to-clone clone-to-clone diff differences:erences: ( a(a)) TheThe measured average average artificial artificial linear linear chromosome chromosome copy copy numbernumber per cell per for cell 11 for clones 11 clones (n = (n8 = for8 for qPCR; qPCR; n = 66 for for retention retention rates), rates), where where the values the values are displayed are displayed as the mean ± standard deviation; (b) MAD2 haploid clones cluster into two phenotypic groups. The as the mean standard deviation; (b) MAD2 haploid clones cluster into two phenotypic groups. graph shows± the measured average artificial linear chromosome copy numbers per cell versus the The graphmeasured shows retention the measured rates, where average values artificial are displayed linear as chromosome the mean ± standard copynumbers deviation; per(c) The cell versus the measured retention rates, where values are displayed as the mean standard deviation; (c) The average artificial linear chromosome copy numbers are significantly different± between the higher averagecopy artificial and lower linear copy chromosome clones (p-value copy < 0.001 numbers denoted areby ***, significantly Student’s t-test); different (d) The between average retention the higher copy and lowerrates copy are significantly clones (p-value different< 0.001between denoted the higher by copy ***, Student’sand lower copyt-test); clones (d) (p The-value average < 0.05 denoted retention rates are significantlyby *, Student’s diff erentt-test).between the higher copy and lower copy clones (p-value < 0.05 denoted by *, Student’s t-test). To characterize clones further, we first measured growth rates by selecting a few clones that have similar retention rates but differed in the average number of artificial linear chromosomes To characterize clones further, we first measured growth rates by selecting a few clones that have (SCSY1108 and SCSY1111), and clones with somewhat similar artificial linear chromosome copy similarnumbers retention but rates different but retention differed ra intes the (SCSY1089 average and number SCSY1008). of artificial However, linear no change chromosomes in growth rates (SCSY1108 and SCSY1111),were observed and relative clones to with wild somewhattype control similarhaploid artificialcells (Table linear S4). We chromosome then analyzed copy 2 wild numbers type but differenthigh retention copy clones rates and (SCSY1089 3 low copy and clones SCSY1008). to determine However, how stable no the change chromosome in growth copy ratesnumbers were and observed relativeretention to wild rates type were control over haploidtime as the cells cultures (Table were S4). propagated We then over analyzed 10 days 2under wild selection type high in CSM- copy clones leu media. Cultures were diluted and renewed each morning in the presence of the leucine

Microorganisms 2020, 8, 1495 7 of 15 and 3 low copy clones to determine how stable the chromosome copy numbers and retention rates were over time as the cultures were propagated over 10 days under selection in CSM-leu media. CulturesMicroorganisms were 2020 diluted, 8, x FOR and PEER renewed REVIEW each morning in the presence of the leucine auxotrophic selection7 of 15 (CSM-leu) and the average artificial linear chromosome copy numbers per cell and retention rates auxotrophic selection (CSM-leu) and the average artificial linear chromosome copy numbers per cell were measured at 0, 3, 5, and 10 days. The 2 high copy clones (SCSY985 and SCSY986) displayed a and retention rates were measured at 0, 3, 5, and 10 days. The 2 high copy clones (SCSY985 and decrease in average copy number per cell over time, while the 3 low copy clones (SCSY1108, SCSY1109, SCSY986) displayed a decrease in average copy number per cell over time, while the 3 low copy and SCSY1110) maintained the same average low copy number over time, where at 10 days there was clones (SCSY1108, SCSY1109, and SCSY1110) maintained the same average low copy number over no significant difference in chromosome copy numbers between the two groups (p-value = 0.1714) time, where at 10 days there was no significant difference in chromosome copy numbers between the (Figure3a). In addition, all clones displayed an increase in retention rates when grown under selection two groups (p-value = 0.1714) (Figure 3a). In addition, all clones displayed an increase in retention for 10 days (Figure3b,c). rates when grown under selection for 10 days (Figure 3b,c).

(a) (b) (c)

FigureFigure 3.3. WildWild typetypeMAD2 MAD2haploid haploid high high copy-low copy-low retention retention rate ra cloneste clones evolved evolved towards towards low copy-highlow copy- retentionhigh retention rate clones rate clones under CSM-leuunder CSM-leu selection: selection: (a) A comparison (a) A comparison of average of chromosomeaverage chromosome copy numbers copy innumbers 2 high copy-lowin 2 high retentioncopy-low rateretentionMAD2 rateclones MAD2 with clones 3 low with copy-high 3 low retentioncopy-high rate retentionMAD2 clonesrate MAD2 over 10clones days over where 10 valuesdays where are displayed values are as thedisplayed mean asstandard the mean deviation. ± standard At deviation. day 10 there At is day no significant10 there is ± dinoff erencesignificant in the difference artificial chromosome in the artificial copy chromosome numbers per cellcopy (p -valuenumbers= 0.1714, per cell Student’s (p-valuet-test); = 0.1714, (b)A comparisonStudent’s t-test); of 2 high (b) A copy-low comparison retention of 2 high rate MAD2copy-lowclones retention with 3 rate low MAD2 copy-high clones retention with 3 rate lowMAD2 copy- cloneshigh retention over 10 days rate whereMAD2 values clones are over displayed 10 days as where the mean valuesstandard are displayed deviation; as the (c) Bothmean populations ± standard ± displayeddeviation; a( significantc) Both populations increase in displayed the retention a significant rates by the increase 10-day in point the relativeretention to therates starting by the point 10-day at daypoint 0 whenrelative grown to the under starting CSM-leu point at selection day 0 when (p-values grown< 0.001 under denoted CSM-leu by selection *** and < (0.01p-values denoted < 0.001 by **denoted respectively, by *** Student’sand < 0.01t -test).denoted by ** respectively, Student’s t-test).

OneOne possiblepossible explanation explanation for for these these trends trends is that is the that artificial the artificial linear chromosomes linear chromosomes were integrating were intointegrating genomic into DNA genomic during DNA the courseduring ofthe the course 10 days, of the where 10 days, the integratedwhere the integrated DNA contained DNA contained the LEU2 auxotrophicthe LEU2 auxotrophic marker. We marker. ruled this We out ruled by screening this out isolatedby screening whole-cell isolated DNA whole-cell for the presence DNA offor both the armspresence of the of artificialboth arms linear of the chromosome artificial linear including chromosome the telomeres including on the one telomeres side, and on also one by side, growing and thesealso by wild growing type clones these inwild the type absence clones of in selection the absence in CSM of selection and measuring in CSM retention and measuring rates (Figure retention S2). Ifrates both (Figure arms of S2). the If artificial both arms linear of chromosomesthe artificial linear are present chromosomes including are a telomere present sequence,including asa telomere detected bysequence, PCR, it wouldas detected indicate by thatPCR, the it artificialwould indicate chromosome that the centromere artificial is chromosome also present, wherecentromere the presence is also ofpresent, an integrated where centromerethe presence would of an create integrated a di-centric centromere chromosome, would whichcreate isa adi-centric lethal event. chromosome, Both arm fragmentswhich is a werelethal detected event. Both at day arm 10 infragments all 5 clones, were and detected no PCR at products day 10 werein all produced 5 clones, byand wild no typePCR controlproducts cells, were ruling produced out genomic by wild integration.type control Incells, addition, ruling whenout genomic clones integration. were grown In in addition, the absence when of leucineclones were auxotrophic grown in selection the absence in CSM, of leucine the retention auxotrophic rates selection in all 5 clones in CSM, fell the to retention below 10% rates by 4in days all 5 indicatingclones fell to that below in wild 10% type by 4 cells days artificial indicating linear that chromosomes in wild type cells were artificial lost when linear the chromosomes selective pressure were oflost the when auxotrophic the selective marker pressure was removed,of the auxotrophic which would marker be was thecase removed, if the LEU2whichauxotrophic would be the marker case if hadthe integratedLEU2 auxotrophic into the genomemarker (Figurehad integrated S2). These into analyses, the genome in combination, (Figure S2). suggest These that analyses, genetically in identicalcombination, wild typesuggestMAD2 thathaploid genetically cells identical display clone-to-clone wild type MAD2 differences haploid in cells the averagedisplay copyclone-to-clone numbers ofdifferences artificial linearin the chromosomesaverage copy thatnumbers they of carry, artific butial that linear these chromosomes differences collapsethat they into carry, a low but copy that number-highthese differences retention collapse rate into state a low when copy cells number-h are propagatedigh retention over timerate state under when selection. cells are propagated over time under selection.

3.2. Mutation in mad2∆ Suppresses Clone-to-Clone Differences in Haploid Cells and Increases Retention Rates Even in the Absence of Selection

Microorganisms 2020, 8, 1495 8 of 15

3.2. Mutation in mad2∆ Suppresses Clone-to-Clone Differences in Haploid Cells and Increases Retention Rates MicroorganismsEven in the Absence 2020, 8, xof FOR Selection PEER REVIEW 8 of 15 Having uncovered the unexpected phenotype of clone-to-clone differences in wild type cells, we Having uncovered the unexpected phenotype of clone-to-clone differences in wild type cells, we next investigated a set of 12 mutant mad2∆ haploid clones to determine the consequences of removing next investigated a set of 12 mutant mad2∆ haploid clones to determine the consequences of removing spindle checkpoint function on the artificial linear chromosomes, where an additional 11 MAD2 spindle checkpoint function on the artificial linear chromosomes, where an additional 11 MAD2 haploid clones were also included in these analyses. Mutant mad2∆ haploid clones suppressed the haploid clones were also included in these analyses. Mutant mad2∆ haploid clones suppressed the clone-to-clone differences observed in MAD2 clones, with 10/11 mad2∆ clones clustering together in a clone-to-clone differences observed in MAD2 clones, with 10/11 mad2∆ clones clustering together in single mid-level copy number-high retention rate group, with only a single clone displaying a low a single mid-level copy number-high retention rate group, with only a single clone displaying a low copy number relative to the rest of the clones (Figure4a,b). Relative to the high retention rate-low copy number relative to the rest of the clones (Figure 4a,b). Relative to the high retention rate-low copy MAD2 clones, mad2∆ mutant clones had a significantly higher artificial linear chromosome copy copy MAD2 clones, mad2∆ mutant clones had a significantly higher artificial linear chromosome copy number per cell (p-value < 0.001) (Figure4c). Furthermore, mad2∆ clones displayed a significantly number per cell (p-value < 0.001) (Figure 4c). Furthermore, mad2∆ clones displayed a significantly elevated retention rate relative to the high copy-low retention rate MAD2 clones (p-value < 0.001) elevated retention rate relative to the high copy-low retention rate MAD2 clones (p-value < 0.001) (Figure4d). (Figure 4d).

(a) (b)

FigureFigure 4. 4. MutantMutant mad2mad2∆∆ haploidhaploid clones clones display display suppression suppression of of clone-to-clone clone-to-clone differences: differences: (a (a) )The The measuredmeasured average average artificial artificial linear linear chromosome chromosome copy copy number number per per cell cell versus versus the the measured measured retention retention ratesrates ( (nn == 88 qPCR; qPCR; n = 66 retention rate) wherewhere valuesvalues are are displayed displayed as as the the mean mean standard± standard deviation deviation for ± for12 mad212 mad2∆ haploid∆ haploid clones; clones; (b) A(b direct) A direct comparison comparison between betweenMAD2 MAD2and mad2 and∆ mad2haploid∆ haploid clones ofclones average of averageartificial artificial linear chromosome linear chromosome copy number copy pernumber cell versus per cell the versus measured the measured retention retention rates where rates values where are valuesdisplayed are displayed as the mean as thestandard mean deviation;± standard ( cdeviation;) The average (c) The artificial average linear artificial chromosome linear copychromosome numbers ± copyare significantly numbers are di significantlyfferent between different higher between copy and higher lower copy copy and clones lower of copy wild clones type MAD2 of wildand type the MAD2mad2∆ andmutant the mad2 clones∆ mutant (p-value clones< 0.001 (p denoted-value < 0.001 by ***, denoted Student’s by t***,-test); Student’s (d) The t average-test); (d retention) The average rates retentionare significantly rates are di significantlyfferent between different higher between copy and highermad2 copy∆ mutant and mad2 haploids∆ mutant (p-value haploids< 0.001 (p-value denoted < 0.001by ***, denoted Student’s by ***,t-test), Student’s but not t-test), between but the notmad2 between∆ clones the mad2 and the∆ clones lower and copy the wild lower type copyMAD2 wildclones. type MAD2 clones.

When culturing cells for large-scale production of biomaterial it may be necessary to change from optimal growth media, which includes factors such as optimal growth rates of the cells, optimal conditions to select for the presence of the artificial linear chromosomes, and the production costs of media raw materials, into a medium designed for optimal induction, production and/or yield of the

Microorganisms 2020, 8, 1495 9 of 15

When culturing cells for large-scale production of biomaterial it may be necessary to change from optimal growth media, which includes factors such as optimal growth rates of the cells, optimal Microorganisms 2020, 8, x FOR PEER REVIEW 9 of 15 conditions to select for the presence of the artificial linear chromosomes, and the production costs of media raw materials, into a medium designed for optimal induction, production and/or yield of target biomolecule(s). In such circumstances, it may not be possible to continue selection for the the target biomolecule(s). In such circumstances, it may not be possible to continue selection for artificial linear chromosome auxotrophy. Thus, we tested directly if mad2∆ strains continued to the artificial linear chromosome auxotrophy. Thus, we tested directly if mad2∆ strains continued to display a high retention rate of the artificial linear chromosomes phenotype even in the absence of display a high retention rate of the artificial linear chromosomes phenotype even in the absence of leucine selection. To this end, MAD2 and mad2∆ haploid cells were grown for 3 days in the presence leucine selection. To this end, MAD2 and mad2∆ haploid cells were grown for 3 days in the presence or absence of leucine selection, and chromosome retention rates were measured (Figure 5). In the or absence of leucine selection, and chromosome retention rates were measured (Figure5). In the presence of leucine auxotrophic selection, cells maintained a high retention rate after 3 days, with no presence of leucine auxotrophic selection, cells maintained a high retention rate after 3 days, with no significant difference between the starting point retention rates in either wild type MAD2 or mutant significant difference between the starting point retention rates in either wild type MAD2 or mutant mad2∆ cells (p-value = 0.4036 and 0.7034, respectively). However, in the absence of selection, wild mad2∆ cells (p-value = 0.4036 and 0.7034, respectively). However, in the absence of selection, wild type type MAD2 cell retention rates fell significantly to 27.5%, while mutant mad2∆ cells only fell to 52.5%. MAD2 cell retention rates fell significantly to 27.5%, while mutant mad2∆ cells only fell to 52.5%. Thus, Thus, even in the absence of selection for the artificial linear chromosomes, mad2∆ mutant cells even in the absence of selection for the artificial linear chromosomes, mad2∆ mutant cells maintained a maintained a significantly higher level of retention compared to wild type MAD2 cells (p-value < significantly higher level of retention compared to wild type MAD2 cells (p-value < 0.01). 0.01).

(a) (b)

Figure 5. Mutant mad2∆ cells display a higher retention rate compared to wild type MAD2 cells even in thethe absenceabsence ofof auxotrophicauxotrophic selection:selection: ( (a)) After 3 days of growth in selectiveselective CSM-leuCSM-leu media no significantsignificant change in retention rates were observed in either wild type MAD2 or mutant mad2∆ cells; ((b)) ByBy contrast,contrast, inin cellscells growngrown forfor 33 daysdays inin CSMCSM inin thethe absenceabsence ofof selection,selection, mad2∆ cells displayed a significantlysignificantly higher retention rate forfor the artificialartificial linearlinear chromosomeschromosomes ((p-value-value < 0.01 denoted by **, Student’s tt-test).-test).

3.3. Analysis of Diploid Wild Type MAD2 MAD2 and Mutant MAD2 mad2 and mad2 mad2 Clones 3.3. Analysis of Diploid Wild Type MAD2/MAD2/ and Mutant MAD2/mad2/ ∆∆ and mad2∆∆/ /mad2∆∆ Clones In studiesstudies of of chromosome chromosome segregation segregation in mitosisin mitosis as well as aswell in industrialas in industrial production production settings, settings, diploid buddingdiploid budding yeast are yeast often are used often as aused tool foras a analyses tool for andanalyses maximal and biomaterialmaximal biomaterial product yield. product Thus, yield. we MAD2 MAD2 extendedThus, we our extended analyses our of the analyses artiﬁcial of linear the chromosomesartificial linear per chromosomes cell into wild typeper cell into/ wildcells type as MAD2 mad2 mad2 mad2 MAD2 MAD2 wellMAD2/MAD2 as mutant cells as/ well∆ asand mutant mutant MAD2/mad2∆/ ∆∆ cells.and mutant Wild type mad2∆/mad2/ ∆ cells.cells Wild displayed type theMAD2/MAD2 clone-to-clone cells di displayedﬀerence in the a pattern clone-to-clone similarto difference wild type in haploids, a pattern yielding similar to two wild phenotypes—one type haploids, whereyielding clones two phenotypes—one contained high copy where numbers clones and contai lowned retention high copy rates, numbers and a second and low set retention of clones rates, with lowand copya second numbers set of and clones high with retention low ratescopy (Figurenumbers6a). and In thesehigh analyses,retention onerates clone (Figure (SCSY1057) 6a). In these was foundanalyses, to haveone undergoneclone (SCSY1057) an integration was found event to that have probably undergone occurred an integration shortly after event transformation that probably and beforeoccurred the shortly clone was after frozen transformation to make a permanentand before stock,the clone where was the frozen chromosome to make wasa permanent maintained stock, at a highwhere level the evenchromosome after 3 days was ofmaintained growth in at non-selective a high level CSMeven after media 3 -days thus, of SCSY1057 growth in wasnon-selective excluded CSM media - thus, SCSY1057 was excluded from further analysis (marked with an arrow, Figure 6a). One notable difference between wild type haploids and diploids is that the average number of artificial chromosomes in the diploid high copy group (4.54 ± 1.16) is significantly less than that compared to haploid wild type clones (6.82 ± 1.70) (p-value < 0.01, Student’s t-test). The 10 MAD2/mad2∆ heterozygous mutant clones displayed a unique distribution: all 10 clones had low

Microorganisms 2020, 8, 1495 10 of 15 from further analysis (marked with an arrow, Figure6a). One notable di fference between wild type haploids and diploids is that the average number of artificial chromosomes in the diploid high copy groupMicroorganisms (4.54 20201.16), 8,is x FOR significantly PEER REVIEW less than that compared to haploid wild type clones (6.82 101.70) of 15 ± ± (p-value < 0.01, Student’s t-test). The 10 MAD2/mad2∆ heterozygous mutant clones displayed a unique retention rates, but the artificial linear chromosome copy numbers fell into 1 of 3 groups, with a low distribution: all 10 clones had low retention rates, but the artificial linear chromosome copy numbers fell copy number (1.743 ± 0.563), mid-level copy number (4.502 ± 1.544) and a single clone with a high into 1 of 3 groups, with a low copy number (1.743 0.563), mid-level copy number (4.502 1.544) and a copy number (9.02 ± 3.27) (Figure 6b). Similar to± mad2∆ clones, mad2∆/mad2∆ clones clustered± into a single clone with a high copy number (9.02 3.27) (Figure6b). Similar to mad2∆ clones, mad2∆/mad2∆ single mid-level copy number-high retention± rate group, where this group also displayed a clones clustered into a single mid-level copy number-high retention rate group, where this group also significant increase in retention rates compared to the high copy number-low retention rate displayed a significant increase in retention rates compared to the high copy number-low retention rate MAD2/MAD2 set of clones (p-value < 0.001, Student’s t-test), even though chromosome copy numbers MAD2/MAD2 set of clones (p-value < 0.001, Student’s t-test), even though chromosome copy numbers between the two groups were not different (p-value = 0.5354, Student’s t-test) (Figure 6c–e). This between the two groups were not different (p-value = 0.5354, Student’s t-test) (Figure6c–e). This display display of an elevated retention rate in the mad2∆/mad2∆ mutants prompted us to test if these clones of an elevated retention rate in the mad2∆/mad2∆ mutants prompted us to test if these clones would also would also display an increased retention rate after 3 days of growth in the absence of selection, like display an increased retention rate after 3 days of growth in the absence of selection, like mad2∆ mutant mad2∆ mutant haploid clones. Compared to both wild type MAD2/MAD2 clones, and MAD2/mad2∆ haploid clones. Compared to both wild type MAD2/MAD2 clones, and MAD2/mad2∆ mutant clones, mutant clones, the mad2∆/mad2∆ mutant clones displayed a significantly higher retention rate (56.0% the mad2∆/mad2∆ mutant clones displayed a significantly higher retention rate (56.0% 5.66%) even in ± 5.66%) even in the absence of selection on CSM after 3 days (p-value < 0.001 and < 0.001,± respectively) the absence of selection on CSM after 3 days (p-value < 0.001 and < 0.001, respectively) (Figure6f). (Figure 6f).

(b) (a)

(d) (c)

Figure 6. Cont.

Microorganisms 2020, 8, 1495 11 of 15 Microorganisms 2020, 8, x FOR PEER REVIEW 11 of 15

(e) (f)

FigureFigure 6. 6. MutantMutant mad2mad2∆∆/mad2/mad2∆∆ diploiddiploid clones clones display display suppression suppression of of clone-to-clone clone-to-clone differences differences and and havehave high high retention retention rates rates even even in in the the absence absence of of selection: selection: ( (aa)) The The measured measured average average artificial artificial linear linear chromosomechromosome copy copy number number per per cell cell ve versusrsus the the measured measured retention retention rates rates ( (n = 88 qPCR; qPCR; nn = 66 retention retention rate) where values are displayed as the mean standard deviation for 10 MAD2/MAD2 genetically rate) where values are displayed as the mean ±± standard deviation for 10 MAD2/MAD2 genetically identicalidentical diploid diploid clones. clones. One One clone clone (SCSY1057; (SCSY1057; mark markeded with with an an arrow) arrow) had had undergone undergone an an integration integration eventevent and and was was excluded excluded from from further further analyses; analyses; ( (bb)) The The measured measured average average arti artificialficial linear linear chromosome chromosome copycopy number number per per cell cell versus versus the the measured measured retention retention rates rates ( (nn = 88 qPCR; qPCR; nn == 66 retention retention rate) rate) where where values are displayed as the mean standard deviation for 9 MAD2/mad2∆ genetically identical diploid values are displayed as the mean± ± standard deviation for 9 MAD2/mad2∆ genetically identical diploidclones. clones. All clones All uniformlyclones uniformly display display a low retention a low retention rate but separatedrate but separated into high into copy, high mid-level copy, mid- copy, leveland lowcopy, copy and groups;low copy (c groups;) The measured (c) The measured average artificialaverage artificial linear chromosome linear chromosome copy number copy number per cell perversus cell theversus measured the meas retentionured retention rates (n = rates8 qPCR; (n =n 8= qPCR;6 retention n = 6 rate) retention where rate) values where are displayed values are as the mean standard deviation for 9 mad2∆/mad2∆ genetically identical diploid clones that suppress displayed ±as the mean ± standard deviation for 9 mad2∆/mad2∆ genetically identical diploid clones thatthe clone-to-clonesuppress the clone-to-clone differences. All differences. clones cluster All clon intoes a cluster single groupinto a withsingle a group mid-level with copy a mid-level number copyand highnumber retention and rates;high (retentiond) A direct rates; comparison (d) A betweendirect comparisonMAD2/MAD2 betweenand mad2 MAD2/MAD2∆/mad2∆ haploid and mad2clones∆/mad2 of average∆ haploid artificial clones linear of average chromosome artificial copy linear number chromosome per cell versus copy the number measured per cell retention versus rates the where values are displayed as the mean standard deviation; (e) the retention rates of artificial linear measured retention rates where values± are displayed as the mean ± standard deviation; (e) the retentionchromosomes rates inof mad2artificial∆/mad2 linear∆ are chromosomes significantly in higher mad2 compared∆/mad2∆ are to wildsignificantly type MAD2 higher/MAD2 comparedhigh copy to clones (p-values 0.001 denoted by ***, Student’s t-test), even though the chromosome copy numbers wild type MAD2/MAD2< high copy clones (p-values < 0.001 denoted by ***, Student’s t-test), even are not different (p-value = 0.5354, Student’s t-test); (f) in cells grown for 3 days in CSM in the absence though the chromosome copy numbers are not different (p-value = 0.5354, Student’s t-test); (f) in cells of selection, mad2∆/mad2∆ cells displayed a significantly higher retention rate for the artificial linear grown for 3 days in CSM in the absence of selection, mad2∆/mad2∆ cells displayed a significantly chromosomes compared to both wild type MAD2/MAD2 and mutant MAD2/mad2∆ clones in a manner higher retention rate for the artificial linear chromosomes compared to both wild type MAD2/MAD2 very similar to mad2∆ haploid clones (p-value < 0.001 and < 0.001 denoted by ***, Student’s t-test). and mutant MAD2/mad2∆ clones in a manner very similar to mad2∆ haploid clones (p-value < 0.001 4. Discussionand < 0.001 denoted by ***, Student’s t-test).

4. DiscussionThese analyses have established that in genetically identical wild type cells there are clone-to-clone differences in artificial short linear chromosome copy numbers per cell and retention rates and that theseThese clonal analyses states are have unstable established over growththat in genetica times oflly 3 days identical in the wild absence type ofcells selection there are and clone-to- 10 days cloneeven underdifferences selection in artificial for the artificialshort linear linear chromosome chromosomes. copy Depending numbers per on thecell experimentaland retention protocol rates and or thatbiomolecule these clonal production states are process, unstable these over clone-to-clone growth times di offferences 3 days couldin the haveabsence a significant of selection impact and 10 on daysthe outcomes. even under selection for the artificial linear chromosomes. Depending on the experimental protocolIn industrialor biomolecule settings, production if the goal process, of a productionthese clone-to-clone scientist differences is to isolate could a large have amount a significant of an impactexogenous on the fragment outcomes. of DNA hosted in the yeast cell as an insert in a short linear artificial chromosome, the yieldIn industrial could be settings, improved if the 3 to goal 4-fold of a by producti using haploidon scientist cells is compared to isolate toa diploidlarge amount cells and of an by exogenousmeasuring fragment the copy numbersof DNA inhosted several in genetically the yeast identicalcell as an clones insert first in by a qPCRshort tolinear select artificial out the chromosome, the yield could be improved 3 to 4-fold by using haploid cells compared to diploid cells and by measuring the copy numbers in several genetically identical clones first by qPCR to select out the high copy-low retention rate clones [1,2]. However, even if an initial screen is performed to work

Microorganisms 2020, 8, 1495 12 of 15 high copy-low retention rate clones [1,2]. However, even if an initial screen is performed to work with a high copy-low retention rate clone, but for biomolecule production cells must be cultured for several days such as during a large volume fermentation run, our results establish that even under selection for the artificial linear chromosome this clonal state will not be stable over an extended period of growth, such as between 3–10 days, and that the clone is likely to evolve towards becoming a low copy-high retention rate population [1,2,16,17]. Thus, in designing a successful high-yield production run the number of days of growth must be minimized. Our results also suggest that if artificial linear chromosomes are to be employed to produce a biomaterial, where in the final stages of production it is necessary to remove media that is optimal for the growth of the cells and the selection of the artificial chromosome and replace it with a media that is optimal for induction or production of the biomolecules, but does not contain the auxotrophic selection, counter-intuitively, it may be better to employ a mad2∆ or mad2∆/mad2∆ strain which might have an improved yield based on a higher copy number and higher retention rates even over 3 days in the absence of selection. This suggestion could be tested, for example, by quantitative Western blots against a protein(s) carried only on the artificial linear chromosomes under the control of an inducible promoter that is not under the control of any negative endogenous feedback loop, and comparing directly the protein yield of a high copy clone versus a low copy clone in MAD2 cells, and/or by comparing the protein yield from a mad2∆ clone versus a wild type MAD2 clone after 3-days growth under non-selective media followed by induction of protein expression. The cellular and molecular mechanism of mad2∆ suppression of the clone-to-clone differences resulting in higher retention rates remains unknown, but in this regard past research observations may be informative. Artificial short linear chromosomes are proposed to present the mitotic spindle checkpoint with a unique challenge. The lengths of the arms on the chromosomes are short enough that when artificial linear paired sister chromatids experience the mitotic spindle pulling forces at their kinetochores in prophase and metaphase they are sometimes pulled apart prematurely before the metaphase-anaphase transition [12,18]. This pre-mature separation is proposed to have several consequences: (i) the artificial chromosomes cannot maintain attachments to the mitotic spindle, where stable attachments require tension between sister chromatid pairs; (ii) this loss of attachment continuously activates the mitotic spindle checkpoint inducing a cell cycle delay until cells decide to proceed in the cell cycle by “mitotic slippage”; (iii) the unattached artificial chromosomes tend to remain in the mother cells because the mother-bud neck creates a physical barrier against the diffusion of the unattached chromosomes into the daughter cell, which would be reflected as a low retention rate [12,18–20]. However, the interpretation of some of these past results may now be confounded by the fact that the number of artificial chromosomes were not measured, so it is unknown for the clones that were employed in these studies if they were high copy-low retention rate clones or low copy-high retention rate clones, where we have observed significant clone-to-clone differences even in genetically identical wild type MAD2 cells. In our current study, for MAD2 haploid cells, the observed clustering of the clones into two groups gives the appearance of a potential non-random distribution. The number of clones analyzed, 21, is not sufficient to determine the exact nature of the distribution of the clones, but we may have sufficient numbers to reject a simple random distribution. If we assume we are not missing a substantial population of clones that have more chromosomes beyond the upper limit we have observed here, namely, about 8 chromosomes for MAD2 wild type haploid cells, then we can apply the X-squared test (Chi-squared test), where a random equal distribution in chromosome copy numbers among yeast transformants would serve as the null hypothesis. If the measured average chromosome copy numbers are binned together onto integers (by rounding up or down), from our 21 clones, the numbers range from 2–8 chromosomes on average per cell (Figure7). Microorganisms 2020, 8, 1495 13 of 15 Microorganisms 2020, 8, x FOR PEER REVIEW 13 of 15

FigureFigure 7. 7. TheThe distribution distribution in in average average chromosome chromosome copy copy numbers numbers in in wild-type wild-type MAD2MAD2 clonesclones is is not not likelylikely to to be be random: random: The The measured measured average average chromosome chromosome copy copy numbers numbers from from 21 21 wild-type wild-type MAD2MAD2 clonesclones are binnedbinned into into whole whole number number integers integers by roundingby rounding the measuredthe measured value value up or down.up or down. The dashed The dashedhorizontal horizontal line represents line represents the average the average number number of 3 clonesof 3 clones expected expected in eachin each group group if thereif there was was a arandom random distribution distribution of of chromosome chromosome copycopy numbersnumbers amongamong clones,clones, which in the Χ X-squared-squared test test is is the the nullnull hypothesis. hypothesis.

TheThe 2–8 2–8 chromosomes chromosomes on on average average per per cell cell observed observed means means there there are are7 expected 7 expected outcomes outcomes if the if distributionthe distribution is random, is random, where where for each for eachtransformed transformed clone clonewe selected we selected we would we would expect expect it to carry it to betweencarry between 2–8 chromosomes 2–8 chromosomes on average on average per percell celland and it would it would have have an an equal equal probability probability of of being being anywhereanywhere in in that that range. range. By By Χ X-squared,-squared, if if we we made made measurements measurements of of 21 21 clones, clones, the the null-hypothesis null-hypothesis is is thatthat we we would would expect expect to to observe observe 3 3 clones clones of of each each clas classs equally—with equally—with 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, or or 8 8 chromosomes chromosomes onon average average per cell. Our Our experimentally experimentally measured measured Χ X-squared-squared value value is is 28.3. 28.3. In In a a system system with with 7 7 expectedexpected outcomes, outcomes, there there are are 6 6 degrees degrees of of freedom, freedom, where where a a Χ X-squared-squared value value of of 28.3 28.3 is is beyond beyond the the criticalcritical value value necessary necessary as as a a cut-off cut-off toto be be able able to to reject reject the the null null hypothesis hypothesis (a (a cut-off cut-off valuevalue of of 22.458 22.458 correspondscorresponds to to the the pp-value-value 0.001). 0.001). However, However, even even if if the the assumption assumption that that cells cells cannot cannot carry carry more more than than 88 chromosomes chromosomes is is true, true, the the rejection rejection of of the the null null hypothesis does does not not give give us us any information about about whatwhat the the true true distribution distribution might might be be other other than than that that it itis isdifferent different from from random—but, random—but, for for example, example, it doesit does not not differentiate differentiate between between a curvilinear a curvilinear distribu distributiontion versus versus a bimodal a bimodal distribution. distribution. To make To makesuch distinctionssuch distinctions measurements measurements on many on many more moreclones clones would would need needto be tomade. be made. WeWe speculate that that in in wild wild type type MAD2MAD2 oror MAD2/MAD2MAD2/MAD2 cellscells a a transformed transformed clone clone may may start start with with aa low low number number or or a a single single copy copy of of 1 1 artificial artificial linear linear chromosome, chromosome, but but as as the the single single cell cell grew grew on on the the selectiveselective media media after after transformation transformation to to make make the the co colonylony that that we we picked picked to to place place into into a a permanent permanent stockstock experienced experienced a a pre-mature pre-mature separation separation event event very very early early on on during during colony colony formation, formation, then then after after mitoticmitotic division, the mothermother cellcell wouldwould have have contained contained 2 2 copies copies and and the the daughter daughter cell cell none, none, where where the thedaughter daughter would would not not have have grown grown under under the the auxotrophic auxotrophic selection. selection. This This mothermother cellcell may have then then beenbeen a a little little more more likely to experience another pr pre-maturee-mature separation separation event event in in subsequent mitosis mitosis simplysimply because because it it was was carrying carrying more more copies copies of of the artificial artificial chromosome. This This may may lead lead to to a a rapid rapid accumulateaccumulate 3 3 chromosomes, chromosomes, and and then then 4, 4, and and then then 5 5 artificial artificial chromosomes chromosomes by by sequential sequential pre-mature pre-mature separationseparation events—each events—each time time becomi becomingng more more and and more more likely likely to to suffer suffer another another premature premature separation separation eventevent in in the the future future as as it it accumulated accumulated more more and and more more artificial artificial chromosomes chromosomes leading leading to to the the bifurcation bifurcation we observed—some clones with 5–8 chromosomes and others with about 2 chromosomes on average per cell. However, in a competitive population of cells over time, the cells that are accumulating extra

Microorganisms 2020, 8, 1495 14 of 15 we observed—some clones with 5–8 chromosomes and others with about 2 chromosomes on average per cell. However, in a competitive population of cells over time, the cells that are accumulating extra chromosomes may ultimately have a selective disadvantage, perhaps when they are carrying about 8 copies of the artificial linear chromosome, which is the upper limit we observed in wild type haploid cells, or 5 copies in wild type diploid cells, where any clone within the population that maintains a low copy of the artificial chromosomes by chance may win out against these unstable and over-burdened cells that have more than about 8 or 5 chromosomes in haploids or diploids respectively. The highest average copy number per cell we measured was for a single MAD∆/mad2∆ clone, where these heterozygous mutant cells are most notable for their uniformly low retention rates, suggesting they frequently experience mis-segregation events. MAD2∆/mad2∆ cells have been demonstrated to have a unique phenotype where they respond normally to loss of attachment events, but they cannot respond specifically to the loss of tension at kinetochores [12]. Thus, there appears to be a correlation between premature loss of tension on short linear chromosomes, and a low retention rate for failing to be able to respond specifically to this loss of tension. We speculate that this instability could lead to a rapid geometric progression in the accumulation of 1, 2, 4, and then 8 artificial linear chromosomes per cell. By contrast, mad2∆ or mad2∆/mad2∆ cells simply do not have a spindle checkpoint pathway at all and advance rapidly through the mitotic cell cycle—typically at a rate of 10–15 min faster than wild type cells [12]. This accelerated progression through mitosis can lead to a balance in chromosome segregation and mis-segregation events: the cells do not respond to lack of attachment to the mitotic spindle, leading to the accumulation of the artificial linear chromosomes in the mother cell leading to an elevated average chromosome number per cell, but may also not give the mitotic forces enough time to build upon the paired sister chromatids of artificial linear chromosomes to pull them apart prematurely during the truncated prophase, which would lead to rapid and successful chromosome segregation as reflected by higher retention rates. Regardless of the underlying mechanisms, our work has empirically established that both basic research scientists and industrial production scientists should be aware of artificial linear chromosome clone-to-clone differences and also aware of how artificial chromosome copy numbers and retention rates can change over time, even under selection. In this study, we only investigated the budding yeast Saccharomyces cerevisiae, but the mitotic spindle checkpoint pathway is very highly conserved among all eukaryotes, and the results reported here likely to be applicable to any other budding species of fungi, especially those that contain a constriction at the mother-bud neck during chromosome segregation, including many members of the phylum Ascomycota such as Pichia pastoris, Kluyveromyces, Brettanomyces, or S. bayanus.

Supplementary Materials: The following are available online at http://www.mdpi.com/2076-2607/8/10/1495/s1, Figure S1: LEU2 and LEU2-∆16 do not have a difference in LEU2 mRNA expression levels, Figure S2: Classic PCR did not detect any evidence for an integration event by 10 days in any of the 5 clones tested, Table S1: The list of strains used in this study, Table S2: A comparison of performing qPCR 16 or 12 times versus 8 times combined with retention assays performed 6 times to measure artificial chromosome copy numbers per cell, Table S3: All raw data for average artificial chromosome copy numbers per cell measurements, Table S4: Haploid clone growth rates compared to wild type cells. Author Contributions: Conceptualization, S.C.S, Y.-S.D. and L.-I.W.; methodology, S.C.S., L.-I.W., Y.-S.D. and Y.-C.L.; software, S.C.S. and L.-I.W.; formal analysis, S.C.S. and L.-I.W.; investigation, S.C.S., L.-I.W., Y.-S.D. and Y.-C.L.; resources, S.C.S., Y.-S.D. and H.-Y.C.; data curation, S.C.S., L.-I.W., Y.-S.D., Y.-C.L. and H.-Y.C.; writing—original draft preparation, S.C.S. and L.-I.W.; writing—review and editing, S.C.S.; funding acquisition, S.C.S., Y.-S.D. and H.-Y.C. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Chang Gung Memorial Hospital (CGMH), grant numbers CMRPD1H0601, CMRPD1J0121, CMRPD1K0121, and BMRPC59 to S.C.S. Acknowledgments: We kindly thank R.Y.L. Wang for advice and use of the departmental Applied Biosystems qPCR machine. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Microorganisms 2020, 8, 1495 15 of 15

References

1. Goldman, S. Genetic chemistry: Production of non-native compounds in yeast. Curr. Opin. Chem. Biol. 2010, 14, 390–395. [CrossRef] 2. Koszul, R. Beyond the bounds of evolution: Synthetic chromosome. How and what for? Comptes Rendus Biol. 2016, 339, 324–328. [CrossRef] 3. Futcher, B.; Carbon, J. Toxic effects of excess cloned centromeres. Mol. Cell. Biol. 1986, 6, 2213–2222. [CrossRef][PubMed] 4. Tschumper, G.; Carbon, J. Saccharomyces cerevisiae mutants that tolerate centromere plasmids at high copy number. Proc. Natl. Acad. Sci. USA 1987, 84, 7203–7207. [CrossRef][PubMed] 5. Hieter, P.; Mann, C.; Snyder, M.; Davis, R.W. Mitotic stability of yeast chromosomes: A colony color assay that measures nondisjunction and chromosome loss. Cell 1985, 40, 381–392. [PubMed] 6. Murray, A.W.; Szostak, J.W. Construction of artificial chromosomes in yeast. Nature 1983, 305, 189–193. [CrossRef][PubMed] 7. Runge, K.W.; Wellinger, R.J.; Zakian, V.A. Effects of excess centromeres and excess telomeres on chromosome loss rates. Mol. Cell. Biol. 1991, 11, 2919–2928. [CrossRef][PubMed] 8. Li, R.; Murray, A.W. Feedback control of mitosis in budding yeast. Cell 1991, 66, 519–531. [CrossRef] 9. Hoyt, M.A.; Totis, L.; Roberts, B.T. S. Cerevisiae genes required for cell cycle arrest in response to loss of microtubule function. Cell 1991, 66, 507–517. [CrossRef] 10. Musacchio, A.; Salmon, E.D. The spindle-assembly checkpoint in space and time. Nat. Rev. Mol. Cell Biol. 2007, 8, 379–393. [CrossRef][PubMed] 11. Schuyler, S.C.; Wu, Y.F.; Kuan, V.J. The Mad1-Mad2 balancing act—a damaged spindle checkpoint in chromosome instability and cancer. J. Cell Sci. 2012, 125, 4197–4206. [CrossRef][PubMed] 12. Barnhart, E.L.; Dorer, R.K.; Murray, A.W.; Schuyler, S.C. Reduced Mad2 expression keeps relaxed kinetochores from arresting budding yeast in mitosis. Mol. Biol. Cell 2011, 14, 2448–2457. [CrossRef][PubMed] 13. Sherman, F.; Fink, G.; Lawrence, C. Methods in Yeast Genetics; Cold Spring Harbor Press: New York, NY, USA, 1974. 14. Ro, D.K.; Ouellet, M.; Paradise, E.M.; Burd, H.; Eng, D.; Paddon, C.J.; Newman, J.D.; Keasling, J.D. Induction of multiple pleiotropic drug resistance genes in yeast engineered to produce an increased level of anti-malarial drug precursor, artemisinic acid. BMC Biotechnol. 2008, 8, 83. [CrossRef][PubMed] 15. Erhart, E.; Hollenberg, C.P. The presence of a defective LEU2 gene on 2 mu DNA recombinant plasmids of Saccharomyces cerevisiae is responsible for curing and high copy number. J. Bacteriol. 1983, 156, 625–635. [CrossRef] 16. Naesby, M.; Nielson, S.; Nielson, C.; Green, T.; Tange, T.; Simon, E.; Knechtle, P.; Hansson, A.; Schwab, M.; Titiz, O.; et al. Yeast artificial chromosome employed for random assembly of biosynthetic pathways and production of diverse compounds in Saccharomyces cerevisiae. Microb. Cell Factories 2009, 8, 45–56. [CrossRef] 17. Hughes, S.R.; Cox, E.J.; Bang, S.S.; Pinkelman, R.J.; López-Núñez, J.C.; Saha, B.C.; Qureshi, N.; Gibbons, W.R.; Fry, M.R.; Moser, B.R.; et al. Process for Assembly and Transformation into Saccharomyces cerevisiae of a Synthetic Yeast Artificial Chromosome Containing a Multigene Cassette to Express Enzymes That Enhance Xylose Utilization Designed for an Automated Platform. J. Lab. Autom. 2015, 20, 621–635. [CrossRef] 18. Murray, A.W.; Schultes, N.P.; Szostak, J.W. Chromosome length controls mitotic chromosome segregation in yeast. Cell 1986, 45, 529–536. [CrossRef] 19. Stern, B.M.; Murray, A.W. Lack of tension at kinetochores activates the spindle checkpoint in budding yeast. Curr. Biol. 2001, 11, 1462–1467. [CrossRef] 20. Wells, W.A.; Murray, A.W. Aberrantly segregating centromeres activate the spindle assembly checkpoint in budding yeast. J. Cell Biol. 1996, 133, 75–84. [CrossRef][PubMed]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).