Appl Microbiol Biotechnol (1992) 37 : 615-620

Applied M'wrobiology Bioteehnology © Springer-Verlag 1992

Studies on plasmid stability, cell metabolism and superoxide dismutase production by P g k - strains of Saccharomyces cerevisiae M. A. Z. Ayub 1, S. Astolfi-Filho 2, F. Mavituna 1, and S. G. Oliver 3 1 Department of Chemical Engineering, 3 Department of Biochemistry and Applied Molecular Biology, University of Manchester Institute of Science and Technology, PO Box 88, Manchester, M60 1QD, UK 2 Laboratdrio de Biologia Molecular, UnB, Bras~lia, Brazil Received 16 December 1991/Accepted 9 April 1992

Summary. A double mutant s o d l / p g k l strain of Saccharomyces cerevisiae has been constructed in order to investigate the effects of different environmental conditions on yeast physiology, plasmid stability, and superoxide dismutase (SOD) production. Strains were transformed with yeast episomal plasmids (YEp) containing both PGK1 and SOD1 genes and were grown on fermentable carbon sources and under vigorous aeration. Under these conditions, the presence of the PGK1 gene was made essential for growth and both genes were efficiently expressed. However, plasmid-borne PGK1 was found not to increase the stability of YEp vectors in batch cultures of P g k - cells. Paradoxically, plasmid stability increased during the respiratory phase of growth. An investigation of the metabolism of P g k cells demonstrated that these glycolytic pathway mutants do not appreciably metabolize glycerol. Thus Pgk +, plasmid-containing, cells have a selective advantage during the respiratory phase of batch growth since they can utilize both glycerol and ethanol.

Introduction The stability of recombinant plasmids is a significant barrier to the efficient commercial production of enzymes or other proteins by genetically manipulated microorganisms. This problem may be overcome by employing synthetic media that permit selection by complementation of auxotrophic mutations in the host or incorporating resistance genes in the plasmid and including the appropriate drug or heavy metal in the growth medium. Both approaches have draw-backs, in terms of expense or safety, for large-scale industrial practice. The yeast Saccharomyces cerevisiae has proved an excellent model microorganism for the production of both homologous and heterologous proteins and has a number of advantages over prokaryotes (Botstein and Fink 1988; Emr 1990). S. cerevisiae is well-characterised geneCorrespondence to: S. G. Oliver

tically and, in particular, a number of mutants carrying lesions in genes specifying glycolytic enzymes have been isolated (Maitra 1971; Lam and Marmur 1978; Ciriacy and Breitenbach 1979). The cloning of genes identified by these mutations has proved a rich source of powerful promoter sequences that enable high level expression of cloned genes (Holland and Holland 1978; Albert and Kawasaki 1978; Hitzeman et al. 1980; Entian et al. 1984; Aguilera and Zimmermann 1986; Rodicio and Heinisch 1987). Moreover, since glycolytic pathway mutants of yeast are unable to grow on fermentable carbon sources, such as glucose (Fraenkel 1982), it has been proposed that complementation of such mutations by the corresponding cloned gene would enable the problem of plasmid instability to be overcome in a self-selecting system (Piper and Curran 1990). In this work, we examine this proposal using the expression of cloned yeast genes encoding superoxide dismutase (SOD) enzymes (Gregory et al. 1974; Fridovich 1986) as a model system. We provide both genetic and physiological data relating to the use of a cloned PGK1 gene to complement the p g k l - mutation in either a SOD1 or a sodI - background. The utility of such a selfselecting system for large-scale batch production of proteins is discussed.

Materials and methods Yeast strains and transformation. S. cerevisiae BC3 strain (MA Tc~ leu2-3, 11 ura3-52 his3 pgkl::TRP1) was kindly provided by Dr. P. W. Piper (Piper et al. 1988). The derivation of strain AD1 (MATch, sodl::LEU2; pgkl::URA3) is fully documented in Ayub (1991). All yeast transformations were performed using the lithium-acetate technique (Ito et al. 1983). Plasmids. The plasmids constructed to transform the yeast are shown in Fig. 1. These are yeast episomal plasmids (YEp), containing replication sequences of the endogenous 2-gm plasmid of S. cerevisiae (the 2-gm origin of replication, ori, and the STB locus, both required for effective partitioning of plasmids during mitotic cell division; Murray and Cesareni 1986. All recombinant DNA manipulations involved in the assembly of these vectors were performed using standard methodologies (Sambrook et al. 1989).

616

NdeI dIII

~

i

.

•

i

!

.

~

d

i

C

i

.

I

BamHI~ ~---

""

HmdlII "

HindIII 60008900 I) I I~OI"S~000 ]Dp -

Fig. 1. Maps of YEp vectors constructed and used in this work to transform BC3 and AD1 yeast strains

Growthmediaandfermentationconditions.

Basic complete medium (YEP) contained 10 g/l yeast extract (Difco) and 20 g/1 peptone (Difco). Various carbon sources were separately sterilized and added to the medium at varying concentrations: 20 g/1 glucose (YEPD); 20 g/1 ethanol and 30 g/1 glycerol (YEPGE). In experiments to assess glycerol metabolism, the medium contained 20 g/1 of ethanol and glycerol concentrations varying from 2.5 g/1 to 30 g/1. Batch cultures to study cell growth, SOD production and plasmid stability were carried out in a 10-1 nominal volume LH bioreactor (LH Engineering, Stoke Poges, Bucks., UK). The general operating conditions were: 28° C, stirrer speed at 500 rpm, airflow rate at 3.5 1/min, pH controlled to 5.5-5.8, with 0.1 ml/l of antifoam (Mazu DF 800 S, Mazer Chemicals), added to the growth medium. The effects of glucose addition to growing cultures of strain AD1 were studied in a 2.5-1 nominal volume LH bioreactor under the same conditions as described above except for the air-flow rate which was 1 l/min. Shake-flask experiments were performed at 28° C and 300 rpm on a rotatory shaker (New Brunswick, USA). The flasks were filled with medium to a tenth of their volume to allow for good aeration. Inocula for all experiments were prepared by transferring a single yeast colony to a 200ml erlenmeyer flask and following cell growth until the culture reached mid-exponential phase, then the culture was used as an inoculum on a 10% (v/v) basis.

Analyticaltechniques.Growth was followed by measuring the optical density at 600 nm. Dry weight determinations were performed by drying known volumes of samples to constant weight at 90° C.

Plasmid stability was determined by the replica plate method (Lederberg and Lederberg 1952). Cell viability was estimated by the methylene-blue staining technique (Trevors et al. 1983). The total SOD content of cells was assayed by the method of Paoletti et al. (1986); MnSOD activity was assessed by the same method but in the presence of 1 mM KCN, and CuZnSOD activity was obtained by the difference between the two rates. The principle of the method is based on the oxidation of NADH to NAD +, mediated by superoxide radicals (O~). One unit of SOD represents the amount of enzyme required to cause 50% inhibition of the rate of NADH oxidation. Oxygen uptake by fermentor-grown cultures was determined by measuring the 02 concentration in the input and output gas flows using an on-line mass spectrometer (Micromass PC Residual Gas Analyzer; VG Quadropoles, UK). The glucose concentration of samples was determined enzymatically using the glucose oxidase reaction, performed in a Glucose analyzer 2 (Beckman Instruments, USA). The medium filtrate obtained from the samples used for dry weight determinations was employed. Samples were diluted in order to bring glucose concentrations within the operating range (0-4.5 g 1-1) of the analyser. Ethanol concentration in the growth medium was analysed using the same sample filtrate by means of gas chromatography. A Pye-Unicam 204 Gas Chromatograph (Philips, UK), fitted with an LDC/Milton Roy C1-10 Integrator (LDC, Ireland) was employed. The glycerol content of culture filtrates was analysed by following the enzymatic oxidation of NADH to NAD ÷ at 340 nm in

617

a Beckman DU-8 spectrophotometer (Beckman Instruments, USA), using a BoehringePMannheim(FRG) test kit. Chemicals. Restriction enzymes were purchased from either Promega or Pharmacia. All other chemicals were of analytical grade and were bought either from Sigma or BDH.

21

25

18 O u~ = 15

20

o

12

~1. ~

E ~. 6

oo

s

Results and discussion

3

0N

Physiology of aerobic batch cultures of plasmid-bearing strains

•

y'

0

v

i

3

•

[

i

6

9

~

i

i

i

i

i

Y

i

12 15 18 21 24 27 30 33 36 Time (h)

30

In order to investigate the utility of PGKl-bearing plasmids in p g k l - yeast cells as a system for the stable and efficient expression of cloned genes, strains BC3 and AD1 were transformed with the plasmids shown in Fig. I. These transformants were grown in aerobic batch cultures under the conditions described in Materials and methods. The cultures were monitored for biomass formation, plasmid stability, carbon source consumption (Fig. 2), and O~ uptake rates (Table 1). All four of the strains investigated showed a similar physiological pattern, exhausting the glucose in the medium after about 12h and accumulating 8-9g/1 of ethanol in the medium. However, significant differences in the growth rate of the transformants in the first (exponential) phase were observed (Table 1). Strain BC3 harbouring the pYEPU plasmid (which contains both the PGK1 and URA 3 genes) exhibited the highest growth rate. In contrast, the same strain when transformed with pYEPS1 (containing the SOD1 and PGK1 genes) grew at a rate slower than any of the other strains. These differences were even more pronounced in the second (respiratory) phase of growth in which the ethanol was consumed (see Fig. 1). The growth rate [P~(~to~)] on the respiratory substrate for BC3 cells harbouring the pYEPU plasmid was almost twice that observed for the same strain transformed with pYEP, while the transformant bearing pYESP1 failed to grow on ethanol at all (Table 1). The sodl- strain AD1, on the other hand, grew on ethanol at an intermediate rate when transformed with pYEPS1. A plausible explanation for the differences in the behaviour of the BC3 and AD1 strains is that the former host contains wild-type copies of the genes encoding both CuZnSOD and MnSOD in its chromosomes. The expression of the plasmid-borne SOD1 is therefore of no selective advantage and may place a metabolic burden on the host, thus explaining its low growth rate and poor biomass yield. Its complete failure to grow on ethanol might be due to an inhibition of mitochondrial function due to the MnSOD precursor overwhelming the system that transports proteins from the cytosol into the organelle. In AD1, on the other hand, the plasmidborne SOD1 complements the defect in the chromosomal copy of this gene and so represents a selective advantage. This not only allows respiratory growth on ethanol but also results in a slightly improved oxygen uptake rate (Table 1) under these conditions.

klJ

10 •

10-

BC3-sp*

25

~ , ~ - - . - - ~ ~>

¢. 27

20 I>

~o 24 ~21

10

'~ 18 Y. ~15

0

3

6

9

12 15 18 21 24 27 30 Time [h)

o 2

33

44

o

5

27 24~

40

'~ 36 32

21 w o

~I-

7 28 ,~ 24

W

12

9 6

N

20 .0.1 1

o°

3

0 LOS 0

3

6

9

12 15 18 21 24 27 30 33 36 Time (h)

38 -I 10 " , ~

APl-sp*

34

2o 18

3O

16... ~

~27

14~.

g23 ~t

-t~1-

10N

s [1.5b 4!= o.6~

'~is

6 4

1.2

0,9 •

4 0 0.1:,r ' ~ . 0

3

, 6

2 ~ , &, , , , , , , o 9 12 15 18 21 24 27 30 33 36 Time (h)

0.38 o

Fig. 2. Physiological results of batch cultures, in YEPD medium (see Materials and methods), of strains BC3(pgkl, SOD1) and ADl(pgkl, socll) transformed with the various plasmids coding for PGK1 and SODI: - p u +, p Y E P U (URA3, PGK1); - p + , pYEP (PGK1); - s p +, pYESP1 (PGK1, SOD1). Full details of these plasmids are given in Fig. 1

618 Table 1. Kinetic parameters obtained for the various batch cultures Strain

BC3-pu ÷ BC3-p ÷ BC3-sp ÷ AD 1-sp +

/~m (GLU) a

0.402 0.373 0.303 0.340

~rn (ETOH) a

0.085 0.046 -0.064

Yx/s (GLU) b

0.186 0.156 0.136 0.172

Yx/s (ETOH) b

0.750 0.664

Maximal SOD ~

qo2m a

MnSOD

CuZnSOD

38.2 31.3

57.1 72.5

66.02 58.08

--

48.7

23.0

--

0.682

45.6

72.8

76.05

a Maximal specific growth rate ( h - 1 ) b Growth yield (g cell dry weight/g carbon source)

c Superoxide dismutase (units/ml) d Maximal specific oxygen uptake rate (rag O2/g cell per hour)

Superoxide dismutase content of cultures

containing and plasmid-free cells might account for the results. However, no overall loss in cell viability was observed, even late in the batch cycle. In the end, a study of the pattern of carbon substrate utilization by pgkl cells provided a more complete explanation (see below).

Figure 3 shows the profile of SOD production of the cultures. For BC3-pu + and BC3-p+ strains, both containing no SODI gene in the plasmid, the production of SOD was similar, with CuZnSOD amounting to 57-77% of the total enzyme content, whereas for BC3-sp +, MnSOD was produced at higher levels than CuZnSOD. This was the strain that produced the lower levels of total SOD and the only one that showed no enhanced production of the enzyme during growth on ethanol, possibly due to the adverse metabolic effects of gene overexpression. The ADl-sp + strain showed slightly higher levels of production of total SOD than observed for BC3-pu ÷ and BC3-p ÷ and MnSOD accounted for more than 60% of the total enzyme produced at all times, confirming efficient gene expression. Furthermore, the reduction in CuZnSOD suggests that MnSOD overproduction adequately protected the cells against superoxide radicals.

Complementation of the pgkl - lesion and plasmid stability Although the vectors used in this work had a number of different components, all contained the PGK1 gene. It has been suggested that such plasmids should be selfselecting in pgkl - hosts in glucose media, where expression of the plasmid-borne wild-type gene is essential for growth (Piper and Curran 1990). In fact, in all cases, the proportion of plasmid-free cells increased during the exponential phase of growth (Fig. 2). This result suggests that there is no obvious selection against p g k l - segregants. This may be because PGK1 is a highly expressed gene (Chem et al. 1984; Chert and Hitzeman 1987) and so plasmid-free segregants carry over sufficient amounts of the enzyme in their cytoplasm to permit growth for some generations. This phenomenon of "phenotypic lag" is well known in classical genetics. The observation that the rate of plasmid loss is reduced, and even reversed, as ethanol is consumed in the second phase of growth is not easily explained. The rate of growth on ethanol is, of course, slower than that on glucose, but chemostat studies (Bugeja et al. 1989) have demonstrated that plasmid loss in yeast increased at slow growth rates. Differences in viability between plasmid-

Metabolism of glycerol and ethanol by pgkl - yeast cells In an attempt to explain the unexpected result that the stability of PGKl-containing plasmids in p g k l - hosts actually improved following the exhaustion of glucose from the medium, we decided to study carbon metabolism in such cells during the respiratory phase of growth. Analysis of medium constituents in the batch culture of ADl-sp + (i.e., the MnSOD- and Pgk-deficient strain transformed with plasmid pYESP1, carrying both the PGK1 nd SODI genes), had shown that both glycerol and ethanol were released into the medium during the fermentative phase of growth and subsequently consumed during the respiratory phase (Fig. 2). In order to further investigate carbon metabolism in p g k l - cells, a series of experiments was performed with the AD1 strain ( p g k l - sod1 - ) bearing no plasmid. Batch culture kinetics for AD1 grown in YEPGE medium are presented in Fig. 4. As expected, this strain showed a relatively slow growth rate and CuZnSOD production (the only SOD produced by this strain) was lower than in any other experiment reported here. More surprisingly, the growth of this strain appeared to be entirely at the expense of ethanol; no glycerol utilisation was observed. These data suggest that p g k l - cells are unable to metabolise glycerol as a source of either energy or carbon. The generality of this effect was confirmed by observations on the p g k l - SOD1 ÷ strain, BC3. Neither strain was able to grow on glycerol alone and the supplementation of ethanol medium with glycerol affected neither the rate nor the extent of biomass accumulation (data not shown). These results, therefore, offer an explanation of the improved stability of PGKl-containing plasmids in p g k I - hosts during the respiratory phase of growth. Plasmid-containing cells would be selected over their plasmid-minus derivatives since only the former are able to use both respirable carbon sources available in the batch growth medium.

619 60 J" , . e . , I v " •

90 80 * 70

/

2

///--'''----~

"

'O,

Ci o

50

~

o

30

N

~r

~

J

J

o

3

6

9

J

I

J

I

I

~

I

I

0

30

~

"5

15,.~

O 24 ~4

12 h5 0.1,

12

6

u

3 0

{/)

0.01

< 0

807°a~.

81

- 34 =

32 "~

~

I.-

10~8°3-~p"~

/ -

/

12 15 18 21 24 27 30 33 36

Time (h)

v // /

g -6

20 ~4 10 o

o.°4

/

,,.,,

40 o ~=

,'~

- 36

~

48

=

¢3

~

ADI(GE) •

"-. 36

,~1

10

0

15

30

45

60 75 Time (h)

90

105 120

Fig. 4. Physiological results and SOD production of strain AD1, growing on YEPGE (see Materials and methods). The only SOD metalloenzyme detected was CuZnSOD. The kinetic parameters observed for this strain were: am = 0.177 h - 1, Yx/s(ETOH)= 0.600 g g - 1, qo2m= 80.34 mg g - 1 h - 1 (for symbols, see Table I)

r,-

Conclusions 0.1

0 3

6

9

12 15 18 21 24 27 30 33 Time (h)

BC3-p÷

~

O "~ "o

100'~ a

~'e ''-'Q-~

--..

60 :~

•- -

N

20 0.1 0.05

. 0

3

. 6

.

.

.

.

. . 0 9 12 15 18 21 24 27 30 33 36 Time (h}

10: ADI_sp+

. ~

o Ch o I..-

115.~

j o'O~

101 ,. 86

1

/

58

The results presented above do not support the hypothesis that PGKl-bearing plasmids in p g k l - hosts represent a self-selecting system for plasmid maintenance. D a t a obtained by others (Piper and C u r r a n 1990) suggesting the utility o f such a system were carried out in continuous culture. In the batch culture experiments reported here, it appears that the phenotypic lag which occurs before the p g k l - genotype can express itself, together with the high viability of p g k l - cells, even under glucose-repressed conditions, b o t h militate against the selection o f plasmid-containing cells. This did not appear to be the case once glucose had been exhausted, however. Here the increase in the p r o p o r t i o n o f plasmid-bearing (PGK1 +) cells can be attributed to their being able to utilise glycerol while p g k l - cells cannot. The finding that glycerol is not essential for sustaining the growth o f p g k l - cells and that such mutants can effectively grow on ethanol as the sole source o f c a r b o n is in agreement with the results o f others (Lam and Martour 1977; Ciriacy and Breitenbach 1979). However, our findings can be explained by the fact that S. cerevisiae has a very low permeability to glycerol (Gancedo et al. 1968; Sols et al. 1971). Since this fatty alcohol is excreted f r o m the cells into the m e d i u m during the fermentative phase o f growth, it then becomes useless to the yeast, which can obtain neither c a r b o n nor energy f r o m it.

148 0.1-, 0

,-

,

,

,

,

,

,

,

,

,

,

3

6

9 12 15 18 21 24 27 30 33 36 Time (h)

0

"~

'~

Fig. 3. Superoxide dismutase (SOD) production of the batch cultures in YEPD medium of the transformed strains BC3 and AD1 transformed with various plasmids, as designated in the legend to Fig. 2. Results represent the volumetric SOD activity of the cultures: U, units

References

Aguilera A, Zimmermann FK (1986) Isolation and molecular analysis of the phosphoglucoseisomerase structural gene of Saccharomyces cerevisiae. Mol Gen Genet 202:83-89 Albert T, Kawasaki G (1982) Nucleotide sequence of the triosephosphate isomerase gene of Saccharomyces cerevisiae. J Mol Appl Gen 1:419-434 Ayub MAZ (1991) PhD Thesis, University of Manchester Botstein D, Fink GR (1988) Yeast: an experimental organism for modern biology. Science 240 : 1439-1443

620 Bugeja VC, Kleinman M J, Stanbury PF, Gingold EB (1989) The segregation of the 2p.-based yeast plasmid pJDB248 breaks down under conditions of slow, glucose-limited growth. J Gen Microbiol 135 : 2891-2897 Chen CY, Hitzeman RA (1987) Human, yeast and hybrid 3-phosphoglycerate kinase gene expression in yeast. Nucleic Acids Res 15:643-660 Chen CY, Oppermann H, Hitzeman RA (1984) Homologous versus heterologous gene expression in the yeast Saccharomyces cerevisiae. Nucleic Acids Res 12:8951-8970 Ciriacy M, Breitenbach I (1979) Physiological effects of seven different blocks in glycolysis in Saccharomyces cerevisiae. J Bacteriol 139:152-160 Emr SD (1990) Heterologous gene expression in yeast. Methods Enzymol 185:231-234 Entian KD, Kopetzki E, Fr6hlich KU, Mecke D (1984) Cloning of hexokinase isoenzyme Pi from Saccharomyces cerevisiae: PI transformants confirm the unique role of hexokinase isoenzyme PII for glucose repression in yeast. Mol Gen Genet 198 : 50-54 Fraenkel DG (1981) Carbohydrate metabolism. In: Strathern JN, Jones EW, Broach JR (eds) The molecular biology of yeast Saccharomyces - metabolism and gene expression. Cold Spring Harbour, N. Y. Fridovich I (1986) Biological effects of the superoxide radical. Arch Biochem Biophys 247:1-11 Gancedo C, Gancedo JM, Sols A (1968) Glycerol metabolism in yeast. Eur J Biochem 5 : 165-172 Gopal CV, Broad D, Lloyd D (1989) Bioenergetic consequences of protein overexpression in Saccharomyces cerevisiae. Appl Microbiol Biotechnol 30: 160-165 Gregory EM, Goscin SA, Fridovich I (1974) Superoxide dismutase and oxygen toxicity in a eukaryote. J Bacteriol 117:456-460 Hitzeman RA, Clarke L, Carbon J (1980) Isolation and characterization of the yeast 3-phosphoglycerokinase gene (PGK) by an immunological screening technique. J Biol Chem 255:1207312080 Holland M J, Holland JP (1978) Isolation and identification of yeast messenger ribonucleic acids coding for enolase, glyceraldehyde-3-phosphate dehydrogenase and phosphoglycerate kinase. Biochemistry 17:4900-4907

Ito H, Fukuda Y, Murata K, Kimura A (1983) Transformation of intact yeast cells treated with alkali cations. J Bacteriol 153 : 163-168 Lain KB, Marmur J (1977) Isolation and characterization of Saccharomyces cerevisiae glycolytic pathway mutants. J Bacteriol 130: 746-749 Lederberg J, Lederberg EM (1952) Replica plating and indirect selection of bacterial mutants. J Baeteriol 63 : 39%406 Maitra P (1971) Glucose and fructose metabolism in a phosphoglucoseisomeraseless mutant of Saceharomyces cerevisiae. J Bacteriol 107 : 759-769 Murray JAH, Cesareni G (1986) Functional analysis of the yeast plasmid partition locus STB. EMBO J 5 : 3391-3399 Oura E (1977) Reaction products of yeast fermentations. Process Biochem 4:19-21 Paoletti F, Aldinucci D, Mocali A, Caparrini A (1986) A sensitive spectrophotometric method for the determination of superoxide dismutase activity in tissue extracts. Anal Biochem 154: 536-541 Piper PW, Curran BPG (1990) When a glycolytic gene on a yeast 2gORI-STB plasmid is made essential for growth its expression levels is a major determinant of plasmid copy number. Curr Genet 17 : 119-123 Piper PW, Curran B, Davies MW, Hirst K, Lockheart A, Ogden JE, Stanway CA, Kingsman A J, Kingsman SM (1988) A heat shock element in the phosphoglycerate kinase gene promoter of yeast. Nucleic Acids Res 16:1333-1349 Rodicio R, Heinisch J (1987) Isolation of the yeast phosphoglyceromutase gene and construction of deletion mutants. Mol Gen Genet 205 : 133-140 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning - a laboratory manual, 2nd edn, Cold Spring Harbour Laboratory Press, Cold Spring Harbor, N. Y. Sols A, Gancedo C, De La Fuente G (1971) Energy-yielding metabolism in yeasts. In: Rose AH (ed) Yeasts, vol 2. Academic Press, London, p 271 Trevors JT, Meerick RL, Russel I, Stewart GG (1983) A comparison of methods for assessing yeast viability. Biotechnol Lett 5 : 131-134