Location Effects of a Reporter Gene on Expression Levels and on Native Protein Synthesis in Lactococcus lactis and Saccharomyces cerevisiae

Abstract

The engineering of industrially important genetically modified organisms by the integration of heterologous genes into the chromosome is often the method of choice for several reasons concerned with long-term stability, homogeneous population distribution, and the enabling of selection without the addition of antibiotics. However, integration may disrupt endogenous gene expression, giving rise to increased levels of toxic metabolic byproducts or activating otherwise silent genes. The position of integration of a foreign gene in the chromosome can also influence its expression levels, and this effect will be of relevance in terms of optimizing protein production parameters. In this study, we determine how the random integration of a foreign reporter gene might affect expression levels and assess the use of proteome analysis to investigate possible effects on synthesis of endogenous proteins in two important food-relevant microorganisms, Saccharomyces cerevisiae and Lactococcus lactis. Eleven L. lactis integrants carrying the gusA gene were analyzed, and expression levels were found to vary by a factor of threefold in contrast to expression levels of lacZ in 18 S. cerevisiae integrants, which showed a 14-fold variation. Of relevance to industry is whether any changes in expression levels might occur as a consequence of storage of the modified strains. Here it is also shown that the above differences in expression levels were not significantly affected by storage of frozen cultures over a period of several months. Analysis of the protein composition of the yeast and lactococcal integrant strains by separation on one-dimensional (1D) and 2D gels showed no significant variations in position beyond those observed in control samples.

The effect of the introduction of a foreign gene to engineer food grade organisms is of great importance in safety terms (13, 18, 20), and is often achieved via chromosomal integration (J. R. Shuster, D. Mayer, H. Lee, Abstr. Am. Chem. Soc., vol. 203, p. 1.119, 1992; 22, 26). In this paper, we take a first step towards analyzing the broader effect of the introduction of a gene on the expression of native protein by using the technique of two-dimensional (2D) gel electrophoresis coupled with powerful computational analyses for comparing gels. The two organisms discussed here are widely used throughout the food industry. Lactococcus lactis is used extensively in starter cultures in the manufacture of dairy products (27), and Saccharomyces cerevisiae is used in the baking and brewing industries as well as a host for the synthesis of recombinant proteins (7). The effect of genomic location on expression of an introduced gene is also becoming increasingly important for genetically modified organisms, and studies in this area are sparse for the organisms discussed here.

Studies of the modulation of expression due to position in prokaryotes are limited to two gram-negative organisms, Escherichia coli (3, 41) and Salmonella typhimurium (37). Beckwith analyzed 11 lac translocation strains and found a twofold variation in expression levels between the origin and termination of replication (3); this study was later corroborated by Sousa et al. (41), who found that levels of β-galactosidase activity differed by two- to threefold in response to chromosomal location. Expression levels in S. typhimurium appear to be similar, and Schmid and Roth (37) analyzing 16 Tn10 integrants of a cluster of his operon genes, found a threefold variation in expression levels with the highest levels being proximal to the origin of replication. A further study of S. typhimurium with a supercoiling-sensitive promoter (32) found similar variations in expression levels and showed that these are not due to localized domains of supercoiling but suggested that they are predominantly due to the operative increase in gene dosage associated with regions close to oriC. In the present study, a reporter gene, gusA, previously used as a reporter gene in L. lactis (33) and placed downstream of a medium-strength lactococcal promoter, was located at several sites within the lactococcal genome. Random integration was achieved via a single-sided recombination mechanism (21) stimulated by asymmetrically ligating randomly generated chromosomal restriction fragments into a suicide vector.

In yeast, the effect of reporter gene integration was also studied by analysis of proteins extracted from integrant strains. The use of proteomics to quantify proteins in yeast is now a well-documented research area, and 2D databases are readily available (16). The effect on expression of gene location might be expected to be more complicated and perhaps show greater variation in yeast than in prokaryotes. It is known that transcriptional activity in yeast is affected by heterochromatic DNA, which gives rise to the phenomenon of position effect variegation (for a review, see Loo and Rine, [25] and Tartoff [43]). This reversible gene-silencing effect, so far found at telomeric regions (11) mating-type loci (25), and ribosomal DNA (40), has been found to be under the control of a number of regulatory genes, SIR1 through 4 (silent information regulators) (1), and also GAL11 (42), NAT1, ARG+D1, and RAP1 (39). In contrast, reports of the activation of silent genes in prokaryotes appear to be restricted to a couple of cases in E. coli (48, 51). The greater need for complex regulatory mechanisms in yeast than in prokaryotes also seems to result in a greater variation in promoter strengths, which could influence a downstream heterologous gene. Two studies have compared the strengths of a range of very weak to very strong promoters when used to express a reporter gene on monocopy (centromeric) plasmids. Mumberg et al. (29), using the lacZ gene, found activity to vary by a factor of 10³ fold. An identical result was also found by Nachen et al. (31) when using the gusA gene as a reporter. Studies of native promoter strengths in prokaryotes appear to be restricted to E. coli. Two studies analyzed nine and four promoters and found a 30- and 75-fold variation in strength, respectively (5, 8). Construction of artificial promoters in L. lactis have yielded strengths that are a magnitude higher than those found in E. coli (19). Whether such variation between organisms exists naturally remains to be seen.

In this study, the lacZ gene, an easily assayed reporter gene used in yeast (12, 35) has been coupled to the constitutive S. cerevisiae promoter for the 5-phosporibosyl-1(α) pyrophosphate synthetase gene, PRS3 (6). The nonreplicable vector was then delivered to the site of integration by including homologous restriction fragments from S. cerevisiae genomic DNA. In this way, several distinct integrant strains were generated and differences in reporter expression levels and effects on native protein synthesis were compared.

MATERIALS AND METHODS

Organisms, plasmids, and culture conditions.

L. lactis MG1363 (9) and the EUROFAN reference strain S. cerevisiae FY1679 (50) were obtained from a collection maintained at the Institute of Food Research. E. coli DH5α (14) was used as a plasmid host for vector construction. Culture conditions for E. coli were as described by Sambrook et al. (36) with selection on L agar (24) and, when appropriate, ampicillin (50 μg/ml), erythromycin (500 μg/ml), isopropyl-β-d-thiogalactopyranoside (IPTG) (20 μg/ml), 5-bromo-4-chloro-3-indoyl-β-d-galactopyranoside (X-Gal) (20 μg/ml), and GlcA (50 μg/ml). Lactococcal strains were grown at 30°C on M17 medium (44) with 0.5% glucose added as a carbon source (GM17). Erythromycin at a final concentration of 5 μg/ml was added where necessary for the selection of transformants. S. cerevisiae FY1679 was cultured at 30°C on YEP medium (39) with 2% glucose added as a carbon source. Geneticin at a final concentration of 200 μg/ml was added for the selection of yeast transformants. The lactococcal suicide vector pFI2281 was constructed from pMG36 (45), pBI101 (Clontech Laboratories, Inc.), and pBR322 (2). The yeast vector pFI2282 was constructed from pSS3-9 (S. Sickinger, personal communication) kanMX3, and kanMX4 (47).

Molecular techniques.

Standard techniques were carried out as described by Sambrook et al. (36). Restriction enzymes, T4 DNA ligase, and Klenow polymerase (Promega Corporation) were all used as described by the manufacturers. Digested chromosomal DNA for vector construction was purified and sized on a Chromaspin 1000 column (Clontech Laboratories), which excludes all DNA fragments less than 1 kb.

DNA extractions.

Chromosomal DNA extraction from L. lactis is a modification of the method described by Lewington et al. (24). Cells from a 100-ml overnight culture were resuspended in 600 μl of 0.25 M sucrose and 50 mM Tris-HCl (pH 8.0) and treated with 0.86 mg of lysozyme per ml for 15 min at 37°C. Cells were lysed by adding prewarmed (37°C) 20% sodium dodecyl sulfate (SDS) to a final concentration of 7% and SDS salt precipitated on ice for 30 min with 5M NaCl added to a final concentration of 0.75 M. The supernatant was extracted with phenol-chloroform (1:1) and then chloroform, and the DNA was ethanol precipitated. Genomic DNA extraction from S. cerevisiae is a modification of the method described in Sherman et al. (38). Cells from a 10-ml overnight culture were resuspended in 0.5 ml of S buffer (0.6 M sorbitol, 5 mM EDTA, 0.05 mM Tris-HCl [pH8.0]), and dithiothreitol and RNase were added to final concentrations of 0.02 M and 0.1 mg/ml, respectively. After incubation at 37°C for 30 min, cells were resuspended in 0.5 ml of S buffer, and Novozym 234 (Novo Industries) was added to a final concentration of 10 mg/ml. The cells were left at 37°C for 2 to 3 h and then incubated overnight following the addition of Pronase E (Sigma) and SDS to final concentrations of 0.1 mg/ml and 0.5%, respectively. The lysate was extracted with phenol-chloroform (1:1), and the DNA was precipitated with isopropanol and washed with 70% ethanol.

Transformations.

E. coli transformation was carried out by the hexaminecobalt chloride method (36). Electroporation of L. lactis was carried out with a Gene Pulser apparatus (Bio-Rad Laboratories) following the procedure of Holo and Nes (17). Transformation of S. cerevisiae was performed with lithium acetate according to the method described by Gietz and Woods (10).

Southern blotting and hybridizations.

DNA routinely resolved on 1% agarose gels was transferred onto Hybond N+ (Amersham) membranes by capillary blotting and hybridized against labeled probes produced according to the Amersham ECL protocol. Copy number determinations were performed using a slot blot apparatus, and signals from hybridizations were compared by laser densitometry to those from reference concentrations of DNA.

Protein extractions.

Proteins for electrophoresis were prepared as follows. For lactococcus, proteins were extracted from 100-ml cultures grown to an optical density at 600 nm (OD₆₀₀) of 0.55 to 0.60. Cells were washed in 50 mM sodium acetate (pH 6.0) and resuspended in 1 ml of a lysis buffer prepared according to instructions described in the manual for the Investigator 2D Electrophoresis system (Genomic Solutions, Inc.). The samples were transferred to 5-ml glass bijous, 1 ml of acid-washed 106-mm-diameter glass beads (Sigma) was added, and cells were broken by vortexing for four times, 1 min each time with 1-min iced cooling intervals. After a short interval of microfuging to remove the beads, cell debris was removed by centrifugation at 20,000 × g for 30 min. Aliquots of 20 μl were stored at −70°C. For yeast, 10 ml of log-phase (OD₆₀₀ = 0.6) cells were treated as described above except that cells were crushed using 425 to 600-mm-diameter acid-washed glass beads (Sigma). Protein extractions for β-glucuronidase assays were from 10 ml of log-phase cells (OD₆₀₀ = 0.6) resuspended in 1 ml of GUS buffer (50 mM NaPHO₄, 10 mM β-mercaptoethanol, 1 mM EDTA, 0.1% Triton X-100). Cells were disrupted with 0.75 ml of 106-mm-diameter glass beads on a bead beater (Stratech, Luton, England) for four intervals of 30 s each with 30-s cooling intervals. Cell debris was removed by centrifugation at 15,000 × g for 15 min at 4°C, and 50 μl of the supernatant was used in assays. For β-galactosidase assays of yeast, 50-ml cultures were grown to an OD₆₀₀ of 0.6, harvested, washed, and resuspended in Z buffer (29), and an equal volume of 425- to 600-mm-diameter glass beads was added. Cells were disrupted by vortexing twice for 1 min with 1-min iced cooling intervals, and then debris was cleared by centrifugation for 5 min at 3,000 × g at 4°C. For the assay, 25 μl of supernatant was used.

Assays.

Protein assays were performed according to Bradford (4), using a Bio-Rad protein assay kit. β-Glucuronidase assays were performed on lactococcal log-phase cells grown to an OD₆₀₀ of 0.5 to 0.6 according to the instruction given for the spectrophotometric assay in Wilson et al. (49) and according to R. A. Jefferson (personal communication). For yeast, a β-galactosidase assay was performed by following the instructions described in Rose et al. (35).

Electrophoresis.

Small-scale 1D protein separations were performed on precast 4 to 12% Bis-Tris gels (Novex) and carried out according to the manufacturer's instructions. Gels were stained with Novex colloidal blue stain. For longer separations, 18 by 25 cm, 12 to 14% gels (Amersham Phamacia Biotech) were used. Gels were stained with colloidal blue and/or silver stained (15). For 2D electrophoresis, 1D separations were performed on 11-cm, linear pH4-7 IPG strips or on 18-cm nonlinear pH3-10 strips for yeast proteins (Amersham Pharmacia Biotech). IPG strips were reswelled in a reswelling buffer that also contained the protein sample. “2D electrophoresis using immobilized pH gradients” (Amersham Pharmacia manual). The buffer was adjusted to also contain thiourea to urea at a ratio of 2:7 (34). All other manipulations were as described in the manufacturer's manual. Separations were carried out with an Amersham Pharmacia Multiphor II apparatus cooled with a Multitemp II temperature-controlled water bath. The 2D separations were on precast 8 to 18% gels from Amersham Pharmacia and were carried out according to the manufacturer's instructions. Gels were silver stained and compared with Bio Image protein analysis software from Genomic Solutions, Inc.

RESULTS

Vector construction.

The lactococcal suicide vector pFI2281 was constructed as follows: the lactococcal P32 promoter (46) was isolated as an EcoRI/SmaI fragment from pMG36, and the ends were repaired with T4 polymerase. This fragment was ligated into the SmaI site of pBI101 in the MCS directly upstream of the GUS gene. The p32-GUS cassette was then excised as a HindIII/EcoRI fragment, and the ends were repaired with T4 polymerase. The cassette was ligated into the BamHI site of pBR322, and the Ery^r gene from pUC7e was cloned as a SalI fragment into the SalI site of this plasmid to form pFI2281 (Fig. 1A). A yeast suicide vector was constructed from pSS3-9, a derivative of Yep356R (30) containing the PRS3 promoter cloned into the SmaI site directly upstream of the lacZ promoter. The Amp^r gene, oriC, and the PRS3 promoter were excised on a single AatII/EcoRI fragment, and the ends were repaired with T4 polymerase. The kanMX3 cassette carrying the lacZ gene and Kan^r gene flanked by the promoter and terminator from Ashbya gossypii was excised as a BamHI/NotI fragment from pFA6a-kanMX3, and the ends were repaired with T4 polymerase. The two cassettes were ligated together, direct repeats present on this construct appeared to cause intrarecombinational problems and were eliminated by digesting the plasmid with EcoRI/BgIII and replacing the kanMX3 module with kanMX4 to form pFI2282 (Fig. 1B).

FIG. 1.

(A) Map of pF12281. Amp^r Ery^r, ampicillin and erythromycin resistance determinants; GUS, β-glucuronidase gene from E. coli; P_P32, lactococcal wg2-specific promoter. (B) Map of pF12282. Km^r, kanamycin resistance determinant; P_PRS3, 5-phospho-ribosyl-1(α)pyrophosphate synthetase promoter; P_TEF and T_TEF, promoter and terminator of A. gossypii TEF gene; lacZ, β-galactosidase gene; T_ADH, terminator of S. cerevisiae ADH1 gene.

Construction of integrant strains.

Lactococcal chromosomal DNA was partially digested with Sau3A, and fragments larger than 1 kb were cloned into the unique BamHI site of pFI2281 to form a minilibrary. A pool of 49 clones containing heterologous-sized inserts of DNA were electroporated into MG1363, and Ery^r transformants were picked and checked for integration by hybridization of uncut DNA. No autonomous plasmid bands were detectable. Separate sites of vector integration were distinguished by digesting integrant chromosomal DNA with DraI, an enzyme that cuts directly downstream of the GUS gene, and then comparing fragment sizes by hybridization against a GUS probe. In this way, 11 integrants carrying the GUS gene at separate chromosomal loci were identified (Fig. 2A). S. cerevisiae FY1679 integrants were generated in a similar manner using random Sau3A genomic fragments ligated into the BamHI site of pFI2282. From 46 integrants, comparisons of chromosomal DNA digested with EcoRI/BgIII yielded 18 clones judged to carry the lacZ gene at separate sites (Fig. 2B).

FIG. 2.

(A) DraI digests of L. lactis integrant chromosomal DNA probed with the gusA gene. Lanes 1 to 11, integrant chromosomal DNA (integrant strains 1 to 11). (B) EcoRI/BglII digests of S. cerevisiae integrant chromosomal DNA probed with the lacZ gene. Lanes 1 to 18, integrant chromosomal DNA (integrant strains 1 to 18).

Reporter gene expression levels.

The lactococcal integrants showed a threefold variation in GUS expression levels. Five of 11 of the activities fell between 80 and 100 nmol/min/mg; the lowest value was 67 nmol/min/mg and the highest was 206 nmol/min/mg (Fig. 3A). These results are consistent with levels previously described for other prokaryotes (3, 37, 41). Measurement of lacZ activities in FY1679 integrants revealed a 14-fold difference between the lowest (0.066 U/mg) and highest (0.933 U/mg) levels, with 7 of 18 values falling between 0.05 and 0.2 U/mg (Fig. 3B). These values show a 4.7-fold-broader range of expression in yeast than in L. lactis. It was also shown that this variation in reporter gene activity found between constructs was not significantly affected by long-term storage at −80°C (Fig. 3). The activity levels of the stored strains closely followed the original measurements; the smallest difference was 6% and the greatest was 17%, with an average difference of 11%, which falls within the error bars for the data. In yeast, the average variation was found to be 17%, with the greatest and smallest differences being 30 and 0.5%, respectively. The greatest variation was for construct 18, which showed the lowest activity. Thus, this difference may reflect an extreme of the detection limit for the assay.

FIG. 3.

(A) β-Glucuronidase activities of lactococcal integrant strains. □, initial assay; , after 20 months of storage at −80°C. (B) lacZ activities of S. cerevisiae integrant strains. □, initial assay; , after 11 months of storage at −80°C.

Protein analysis.

Samples from integrant strains were separated on precast 1D gels and protein positions analyzed with commercially available software. Cross-comparison of more than 40 lactococcal proteins revealed one or two extra bands at very low molecular masses (less than 5 kDa); in some samples, however, it was found that the presence and position of these bands were inconsistent and that they most likely represented rare artifacts arising from random proteolytic cleavage, either caused by the isolation procedure itself or enzymatically induced. There were no significant differences to within a 5% margin in the positions of any of the higher-molecular-mass bands (data not shown). With yeast, a comparison of around 60 proteins also revealed some construct-to-construct variation in the presence or absence of a few very-low-molecular-weight bands (less than 10 kDa); again, these differences were inconsistent from gel to gel, and it is suspected that they also represent proteolytic effects. Positional variation between all bands was allowed to be within a 5% margin of error. Greater resolution was achieved by the technique of 2D electrophoresis, so for lactococcus it was possible to separate between 300 to 400 proteins and for yeast around 600 to 800 proteins (Fig 4). For lactococcus and yeast, image analysis revealed a consistency in the relative positions of proteins; however, intensity of some of the fainter spots was variable from gel to gel; this could represent differences in loading or staining inconsistencies (e.g., temperature variations), as well as small changes in protein solubility. It should be noted that 2D electrophoresis is a very sensitive technique in terms of protein preparation, subsequent treatment of the samples, run conditions, and staining protocols and that the most accurate data arises from averages of the spot intensities derived from several repeat gels. Thus, although quite good data in terms of spot position are obtainable from two gels of each sample, differences in spot intensities should be interpreted with caution and require the comparison of at least several identical gels to gain a meaningful result.

FIG. 4.

(A) Example of 2D gel electrophoresis of L. lactis proteins. 1D separations were on 11-cm pH4-7 linear gradient IPG strips; 2D separations were on 8 to 18% SDS gradient gels. (i) Duplicate gels of total MG1363 proteins; (ii) duplicate gels of total protein isolated from integrant 1. (B) Example of 2D gel electrophoresis of S. cerevisiae proteins. 1D separations were on 18-cm pH3-10 nonlinear gradient IPG strips. 2D separations were on 12 to 14% SDS gradient gels. (i) and (ii), duplicate gels of total FY1679 proteins; (iii) and (iv), duplicate gels of integrant 1 proteins. All gels were subjected to computer-assisted analysis as described in Results.

DISCUSSION

A reporter gene was randomly integrated into the yeast and lactococcal genomes by the process of single-crossover recombination to yield a set of constructs that differed from each other in terms of the location of the site of integration. Comparison of the expression of the GUS reporter gene in 11 L. lactis constructs showed a threefold variation in expression with a mean 11% variation following prolonged storage at −80°C. In contrast, expression of the lacZ reporter gene in S. cerevisiae showed a 14-fold variation in expression, and this difference may reflect the greater complexity of the yeast genome in terms of the presence of regions of gene silencing or perhaps differences in promoter strengths between the two organisms; the possibility of read-through expression from a chromosomal promoter cannot be ruled out. The introduction of a stably expressed heterologous gene into yeast would require the avoidance of heterochromatic regions of the genome. Long-term storage of both yeast and lactococcal strains does not significantly affect the expression of reporter genes, indicating that little if any genetic change occurs in frozen stocks. Comparative analysis of the expression of native proteins at the level of 1D gel electrophoresis revealed no differences within the yeast and lactococcal constructs apart from inconsistent changes in very-low-molecular-weight peptides which probably represent proteolytic degradation artifacts. A 2D gel analysis of the constructs resulted in an approximately 10- to 15-fold increase in the number of proteins available for comparison. Although the position of the proteins remained relatively invariable (to within 5%), there was a two- to fivefold variation in the intensities of some of the fainter proteins. This variation was inconsistent from gel to gel and may reflect differences in staining parameters, sample solubility, sample preparation, or storage conditions. Thus, at least several gels of the same sample need to be averaged to establish valid changes in spot intensity. In the future, a microarray analysis should provide a robust complementary method for studying the effects of an integrated foreign gene on genome-wide expression levels.