Current Genetics

Current Genetics (1983) 7 : 1 1 7 - 1 2 2

© Springer-Verlag 1983

DNA Insertions which Affect the Expression of the Yeast Iso-2-Cytochrome c Gene Cynthia F. Wright, Deborah A. Walthall, Jeremy M. Boss, and Richard S. Zitomer Department of BiologicalSciences, State Universityof New York at Albany, Albany, New York 12222, USA

Summary. The plasmid YCpCYC7(2) was constructed containing the Saccharomyces cerevisiae CYC7 gene, encoding the iso-2-cytochrome c protein, replicative sequences and selective markers from both E. coli and yeast, and the centromere of yeast chromosome III. The expression of the plasmid-CYC7 gene in yeast was similar to the low level expression characteristic of the chromosomal CYC7 gene. A number of insertions into the sequences 5' to the gene were constructed in vitro. The insertion at 142 bp 5' to the coding sequence of a 400 bp fragment which lies 5' to the CYC1 gene and is known to be essential for the high rates of CYC1 transcription increased transcription of the CYC7 gene to levels characteristic of CYC1 transcription. On the other hand, the insertion of random DNA fragments at the same position gave mostly decreased CYC7 transcription. In addition to these in vitro constructions, a mutant plasmid was selected which had increased CYC7 transcription. This mutation was caused by the insertion of the bacterial IS1 element 313 bp 5' to the CYC7 coding sequence. The significance of these results is discussed in terms of two alternative models for CYC7 gene expression. w o r d s : Transportable elements - Insertion mutations - Expression mutants - Yeast CYC7 gene Key

Introduction

A growing body of evidence has emerged from studies of the expression of yeast genes which implicates regions

offprint requests to: C. F. Wright

one to two hundred basepairs 5' to the transcription initiation site in the regulation of gene expression (Guarente and Ptaslme 1980; Williamson et al. 1981; Montgomery et al. 1982; Lowry et al. 1983). The existence of such distant upstream regulatory sequences is unprecedented in prokaryotic systems, and how they exert their effect along the DNA is as yet undetermined. We previously reported two mutations involving major DNA rearrangements upstream from the iso-2-cytochrome c gene, CYC7, which caused increased expression of that gene (Montgomery et al. 1982). Such mutations can be selected from a cycl background by the following strategy. Iso-l-cytochrome c, encoded by the CYC1 gene (Sherman et al. 1966), comprises 95% of the cytochrome c in most laboratory strains, and cycl mutants are unable to grow on lactic acid as the sole energy source due to the resulting cytochrome c deficiency. Mutations which increase the expression of the CYC7 gene overcome this deficiency and permit growth on lactic acid (Clavflier et al. 1969). Despite this strong selection, relatively few CYC7-1inked mutations of this type have been isolated, and all those reported involve substantial DNA rearrangements (Errede et al. 1981; McKnight et al. 1981; Montgomery et al. 1982). To increase the number and broaden the range of CYC7 over-expressing mutations which would allow a detailed analysis of the CYC7 transcriptional signals, we have constructed a yeast transforming plasmid containing the CYC7 gene in which mutations can be constructed in vitro and which can be subjected to in vitro mutagenesis. We have inserted a variety of DNA fragments into this plasmid in the region 5' to the CYC7 gene in an attempt to assess what kinds of sequences and at what locations cause increased expression. We have also isolated an over-expressing CYC7 mutation in this plasmid, demonstrating that plasmids can be used to select randomly induced mutations.

118 Materials and Methods

Yeast Strains and Transformation. The following Saecharomyces cerevisiae strains were used: D311-3A, CYC1, CYC7 (Sherman et al. 1974); GM3C-2, cycl, cyc7 (Faye et al. 1981); and Z065 derived from a mating between GM3C-2 and D13-1a (Struhl et al. 1979) with the phenotype ~, trpl-1, leu2-3, Ieu2-112, his4-519, cycl-1, CYC7. All transformations of yeast cells were carried out with GM3C-2 as described. The procedure used was a modification (Lowry et al. 1982) of that described by Hinnen et al. (1978).

Plasmids, Plasmid Constructions and Bacterial Transformation. The following plasmids were used: YRp7 containing pBR322 and the TRPl-arsl sequence (Struhl et al. 1979); pYeCYC7 (0.66) containing pBR322 and a 660 bp fragment containing the CYC7 coding sequence and pYeCYC7(2.2) containing pBR322 and a 2.2 kb fragment containing the CYC7 gene (Montgomery et al. 1980); and pYE(CEN3)41 containing pBR322 and the centromere from yeast chromosome III (Clarke and Carbon 1980). Transformation of the E. coli strain HB101 (Boyer and Roulland-DussoLx 1969) was carried out as described by Cohen et al. (1972). The plasmid YCpCYC7(2) was constructed in the following manner. A 2.2 kb ClaI fragment containing the CYC7 gene was excised from pYeCYC7(2.2) and ligated with ClaI digested and bacterial alkaline phosphatase treated YRp7. The ligation mixture was used to transform E. coli cells, and the resulting transformants were screened by restriction analysis for those containing the correct construction (Birnboim and Doly 1979). The appropriate plasmid was then digested with BamH1, eliminating a 500 bp CYC7-pBR322 segment, treated with bacterial alkaline phosphatase, and ligated with the 2.2 kb BgllI-BamH1 fragment from pYe(CEN3)41 which contains the centromere of chromosome III. The ligation mixture was transformed into E. coli cells and the ampicillin resistant, tetracycline sensitive transformants containing YCpCYC7(2) were identified by restriction analysis. Restriction enzymes and T4 ligase were purchased from BRL with the exception of ClaI which was obtained from BoehringerMannheim Inc. The reaction conditions were those recommended by the vendor.

C.F. Wright et al.: Expression Mutants in Yeast screening performed as described (Messing et al. 1977). A 17nucleotide primer was purchased from Collaborative Research.

Results Expression o f the Plasmid Borne CYC7 Gene The CYC7 containing plasmid, YCpCYC7(2) is shown in Fig. 1A. It contains the yeast 1.4 kb TRPl-arsl sequence (Struhl et al. 1979), pBR322 sequences, the 2.2 kb BglII-BamH1 fragment containing CEN3 (the centromere from yeast chromosome III) (Clarke and Carbon 1980), and a 2 kb ClaI-BamH1 fragment conraining the CYC7 gene with 900 bp o f 3' flanking sequences and 800 bp of 5' flanking sequences (Montgomery et al. 1980). The TRPl-arsl segment allows the selection o f Trp ÷ transformants from trpl auxotrophs and plasmid replication. The CEN3 fragment confers mitotic stability to this plasmid and also holds the copy number at one per cell as reported b y Clarke and Carbon (1980) and confirmed b y us using their procedures. The expression o f the CYC7 gene on the plasmid was phenotypically similar to that o f the chromosomal gene (Fig. 1). Wildtype yeast cells are capable o f growth on a variety of non-fermentable energy sources inchiding glycerol and lactic acid as shown in Fig. lB. Cells deficient in iso-l-cytochrome c, cyel, are still capable of growth on glycerol due to the residual 5% cytochrome c in the form o f iso-2, but are incapable o f growth on lactic acid. Cells completely deficient in cytochrome c, cycl, cyc7, are, o f course, incapable o f growth on either energy source. As can be seen in this figure, when cycl, cyc7 cells were transformed with YCpCYC7(2), the transformants showed a pattern o f growth on glycerol and lactic acid identical to that o f cycl, CYC7 cells.

Northern Analysis. Cells were grown on either raffinose (derepressed) or glucose (repressed) medium as described by Lowry et al. (1983). Total cellular RNA was prepared as described by Zitomer and Hall (1976) and size-fractionated by electrophoresis in agarose-formaldehyde gels (Rave et al. 1979). The RNA was blotted ontogenescreenfilters (New England Nuclear) as described by Thomas (1980). Hybridization was carried out as described by Lowry et al. (1983) using nick-translated (Maniatis et al. 1975) pYeCYC7(0.66) as the probe.

DNA Sequence Determinations. DNA sequence determinations were carried out using the dideoxynucleotide chain termination method of Sanger et al. (1977) except that reverse transcriptase was substituted for DNA polymerase. The DNA was subcloned into the single stranded DNA vector M13mp8, obtained from J. Messing through BRL. The XhoI-BamH1 fragment containing the IS1 insert was subcloned into Sall-BamH1 digested M13rap8. In addition, the PstI fragment extending from the PstI site within the CYC7 segment (-287) to the PstI site of IS1 was subcloned into the PstI site of M13mp8 in the orientation that permitted sequence determination from the CYC7 PstI site into IS1. M13mp8 was prepared, transfections carried out, and

The CYC1 Transcriptional Sequence The CYC7 gene is normally transcribed at an extremely low rate (Montgomery et al. 1982) which might be due to a weak transcription start site, a weak upstream regulatory sequence, or a combination o f both. To determine whether a new upstream sequence could increase gene expression, we inserted a 400 bp sequence flanked b y XhoI sites which lies 250 bp 5' to the coding sequence o f the CYC1 gene. This sequence is known to contain an upstream sequence for the CYC1 gene which is transcribed at twenty-times the level o f CYC7 (Zitomer et al. 1979; Guarente and Ptashne 1980; Lowry et al. 1983). This XhoI CYC1 fragment was inserted into YCpCYCT(2) at the unique XhoI site at -142 in the same orientation as that in the CYC1 gene (determined b y restriction mapping). Yeast transformed with this

119

C. F. Wright et al.: Expression Mutants in Yeast /~

Pst I

EcoR1

Fig. 1A and B. Diagram of the plasmid YCpCYC7(2) and its expression in yeast. A The plasmid YCpCYC7(2) with the relevant restriction sites. Nucleotides are numbered with the A in the CYC7 ATG initiation codon as 1, and nucleotides 3' in positive integers and those 5' in negative integers. B The S. cerevisiae strains used were as follows: CYC1, CYC7, D311-3A; cycl, CYC7, Z065; cycl, cyc7, GM3C-2

plasmid, YCpCYC7(2)X, grew well on lactic acid plates indicating increased expression o f the CYC7 gene. Northern blot analysis of RNA isolated from these transformants demonstrated that increased expression o f the gene was due to increased levels of m R N A (Fig. 2). The length of the CYC7 m R N A was unchanged by the insertion of the 400 bp XhoI fragment suggesting that the transcriptional initiation site at 77 nucleotides 5' to the initiation codon (Montgomery et al. 1982) was unaltered. In addition glucose, which represses CYC1 gene transcription (Zitomer et al. 1979), repressed CYC7 m R N A levels in cells transformed with either the wildtype or the insertion plasmid. We previously reported that in other strains the expression of the wildtype CYC7 gene was not glucose repressible (Montgomery et al. 1982). This difference is probably due to the variation in catabolite repression in different strains. Two larger transcripts also appeared on this blot. These represent plasmid specific transcripts homologous to pBR322 and unrelated to the CYC7 gene (Breunig Fig. 2A-C. Autoradiograph of the northern blot analyses of RNA levels in transformed cells. Total cellular RNA was prepared from GM3C-2 ceils transformed with YCpCYC7(2), A YCpCYC7(2)X, B or YCpCYC7(2)IS1, C and grown under either repressed (2% glucose, G) or derepressed (2% raffinose, R) conditions. The arrow indicates the position of the CYC7mRNA

120

C.F. Wright et al. : Expression Mutants in Yeast -375 ATGACAAAGTCATCG

-360 GGCATTATCTGAACATAAAA

-340 CACTATCAATAAGTTGGAGT

-320 CATTACC

--IS1

CYC7 CGTGCCTTCTCTG CYP3-15 CYP3-4 -300 I -280 I -260 AGAAGGGTCTGCAGTCCCCC GCCGAGGGGTCTTTTCCCAC CTTCTCAAAGCTAATAGCGA -240 TAATAGCGAGGGCATTTATT

-220 CAACTTCCAACTACTATAAG

-200 TGGCCGCAAGGGGCAAAGAC

-180 AAAGGCACACAACATATATA

-160 TATATCGTGTTGTGAAGCTC

-140 GAGAAGATTAGATCAGAATA

-120 GTTCTCTTTTTGTTGAGGTT

-i00 GAAACAAAATCAAAGACTTA

-80 TACAAGAAGATCACATACAA

-40

-20

Lap s i t e -6O GCATTTATTCACATTACTTT

AAGTAAACTTCAGTAAACTA

CATTACATCATAAACAAAAC

+i ATG GCT AAA GAA AGT ACG

Fig. 3. DNA sequence of a portion of the YCpCYC7(2)-IS1 plasmid. The sequence from the XhoI site (-142) to -375 was determined as described in Materials and Methods. The sequence from -142 to +18 was taken from Montgomery et al. (1980). From -142 to -313 the sequence is identical to that previously determined for CYC7 (Montgomery et al. 1980 and D. L. Montgomery personal commun.) and from -314 to -375 with that for IS1 (Ohtsubo and Ohtsubo 1978)

et al. 1982; P. E. Clark and R. S. Zitomer, unpublished results).

upstream sequence may be necessary to mimic the CYC7 transcription signals.

Random Fragment Insertions

Selection of CYC7 0ver-Expressors

To determine whether any sequence inserted at -142 5' to the CYC7 coding sequence would cause increased expression, yeast DNA was digested with XhoI plus Sail and inserted into the XhoI site of YCpCYC7(2). The resulting plasmids were transformed into E. coli, and the transformants were screened for plasmids containing inserts. Six plasmids containing different inserts ranging from 0.8 to 16.7 kb were each transformed into yeast, and the Trp + transformants were screened for growth on glycerol and lactic acid plates. None of the transformants showed increased expression of the CYC7 gene as determined by their inability to grow on lactic acid plates. Three (insert sizes, 0.8, 1.7 and 16.7 kb) did not grow on glycerol plates indicating very low expression of the CYC7 gene. One (insert size, 7.3 kb) grew poorly on glycerol plates, and two (insert sizes, 8.1 and 15.3) grew as well as transformants containing the parental plasmid. These results indicate that only an authentic upstream sequence such as the CYC1 XhoI sequence can cause increased expression of the CYC7 gene when inserted into the XhoI site. In addition, because most insertions into the XhoI site decreased CYC7 expression, we concluded that there is a positive upstream sequence 5' to this site which was removed upon introduction of new sequences. In the two cases in which expression was similar to wildtype, it is possible that a sequence bearing some homology to upstream sequences was located on the inserted fragment. Since the CYC7 gene is expressed at such low levels, only weak homology with an ideal

Mutants which overexpress the CYC7 gene can be obtained from cycl CYC7 cells by selection for growth on lactic acid plates. Despite a number of mutations obtained in this manner, no cis-dornlnant, CYC7-1inked point mutations have been found. In an attempt to isolate such mutations, the YCpCYC7(2) plasmid was mutagenlzed in vitro with hydroxylamine. After the mutant pool was amplified in E. coli, yeast cells were transformed, and lactic acid plus transformants were selected. To our surprise, the only transformant which grew on lactic acid carried a plasmid with a 750 bp insert between the PstI site (-287) and the BamH1 site (-800). The increased expression of the CYC7 gene was visualized by Northern blot analysis (Fig. 2). The transformants containing the mutant plasmid, YCpCYC7(2)-IS1, had higher levels of CYC7 mRNA than did transformants containing the wildtype plasmid, and this higher level was subject to glucose repression. The nature of the inserted sequence was initially investigated by Southern analysis using YCpCYC7(2)IS1 as a probe for both yeast and E. coli genomic DNA. The results (not shown) indicated that the plasmid contained sequences repeated in the E. coli genome, suggesting that the insert was a repetitive mobile element, probably IS1, a 768 bp insertion dement of E. coli (Ohtsubo and Ohtsubo 1978). Our suspicions were confirmed by sequence analysis which showed that the insert integrated at -313 and was identical to IS1 (Fig. 3).

121

C. F. Wright et al. : Expression M u t a n t s in Yeast

Because this mutation was found in a hydroxylamine mutagenized plasmid population, it was necessary to determine whether the ISl sequence or a point mutation outside the sequenced region caused the lactic acid plus phenotype. To resolve this question the XhoI-BamH1 fragment containing the insert was excised from the mutant plasmid and ligated into XhoI-BamH1 digested wildtype plasmid. The new plasmids were then transformed into yeast and found to give lactic acid plus transformants, indicating that the mutant phenotype resulted from IS1 insertion.

Discussion

The results reported here demonstrate that the expression of the CYC7 gene from the centromere-containing plasmid YCpCYC7(2) was similar to that of the chromosomal gene. This similarity of expression and the stability of the plasmid allowed the identification of transcriptional mutations in the plasmid-borne gene by the altered phenotypes of the transformants; this was not always true for the noncentromeric versions of this plasmid (unpublished results). In particular, the isolation of the IS1 insertion mutation, although clearly not the hydroxylamine point mutations we desired, did demonstrate that this type of plasmid can be used to select among randomly induced mutations those rare ones which give the desired altered phenotype. This is an extremely important point to establish if the plasmid-expression system in yeast is to reach its full potential. The use of random mutagenesis of plasmids can quicldy pin-point those regions which are most important to gene expression and which should be the target of the more powerful site specific mutagenesis techniques. In this light, our failure to obtain point mutations was disappointing. We met with the same lack of success in attempts to obtain hydroxylamine-induced point mutations affecting transcription of the CYC1 gene (C. V. Lowry and R. S. Zitomer, unpublished results). We can only conclude that such mutations are extremely rare, if possible, and that more drastic sequence changes are required to obtain transcriptional mutations in these genes. Fortunately, preliminary results with the plasmid-bome CYC7 gene indicate that the insertion of synthetic octanucleotides can cause transcriptional changes manifest in altered cell phenotypes. We have reported here the construction of a number of insertion mutations in the CYC7 gene at -142. Insertions of random sequences of DNA at this position either negatively affected transcription or left transcription unaltered. On the other hand, the insertion of an authentic transcriptional stimulatory sequence, the CYC1 upstream sequence, caused increased CYC7 expression. We interpret these results as suggesting that

a positive transcriptional sequence lies 5' to the -142 site, and that the random insertions separate this sequence from the CYC7 protein coding region. Only its replacement with a stronger upstream sequence, like that of CYC1 can yield higher levels of CYC7 expression. The mechanism by which IS1 insertion causes increased expression of the CYC7 gene must be considered in conjunction with the two other cis-dominant mutations previously reported to increase CYC7 transcription: CYP3-4, a Tyl element insertion at -269, and CYP3-15, an inversion or deletion with one endpoint at-285 (Montgomery et al. 1982). In addition any model must include our above conclusion that there is an upstream sequence 5' to -142 which is required for CYC7 transcription. We propose two alternative models which differ in that the first assumes these mutations placed special sequences next to the CYC7 gene, while the second assumes a negative regulatory element was removed by these otherwise neutral inserted sequences. The first explanation assumes that the normal upstream sequence of the CYC7 gene is a weak one, and that each of the three DNA rearrangements fused a new powerful positive regulatory sequence to the gene. This hypothesis can be readily accepted for the Tyl element insertion since this mutation also places the CYC7 gene under the same mating type control that regulates the Tyl transcripts (Clavllier et al. 1969, 1976). However, it is less obvious that the IS1 and CYP3-15 mutations placed strong positive regulatory sequences 5' to the CYC7 gene. The alternative explanation invokes a negative regulatory element which lies 5' to a strong positive sequence and these two sequences acting antagonistically give the low wildtype level of CYC7 transcription. The increased expression in the three mutations would have resulted from either the deletion of this negative regulatory element (CYP3-15) or its removal by insertion of DNA between it and the positive regulatory sequence (CYP3-4 and IS1). This hypothesis requires that the positive regulatory sequence lie between -142 (based upon the random insertion experiments described above) and -269 (the site of the Tyl insertion). The negative regulatory element must lie 5' to -313 (the site of the IS1 insertion). Experiments are in progress to distinguish between these two models and suggest the latter, more complex one, is correct. Whichever proves correct, it is clear that the transcription of this gene is governed by sequences well upstream from the mRNA CAP site.

Acknowledgement. This

work was supported by an N.S.F. grant to R. S. Z., an N.I.H. RCDA to R. S. Z., and an N.S.F. predoctotal fellowship to C. F. W, We t h a n k M. A. Zitomer for the tetrad analyses o f transformed cells.

122

References Birnboim HC, Doly J (1979) Nucleic Acid Res 7:1513-1523 Boyer HW, Roulland-Dussoix D (1969) J Mol Bio141:459-472 Breunig KD, Mackedonski V, Hollenberg CP (1982) Gene (in press) Clarke L, Carbon J (1980) Nature 287:504-509 Clavllier L, Pere G, Slonimski PP (1969) Mol Gen Genet 104: 195-218 Clavilier L, Pere-Aubert G, Somlo M, Slonimski PP (1976) Biochimie 58:155-172 Cohen SN, Chang ACY, Hsu CI (1972) Proc Natl Acad Sci USA 69:992-999 Errede B, Cardillo TS, Sherman F, Dubois E, Deschamps J, Wiame JM (1981) Cell 22:427-436 Faye G, Leung DW, Tatchell K, Hall BD, Smith M (1981) Proc Natl Acad Sci USA 78:2258-2262 Guarente L, Ptashne M (1980) Proc Natl Acad Sci USA 78: 2188-2192 Hinnen A, Hicks JB, Fink GR (1978) Proc Natl Acad Sci USA 75:1929-1933 Lowry CV, Weiss JL, Walthall DA, Zitomer RS (1983) Proc Natl Acad Sci USA 80:151-155 Maniatis T, Jeffrey A, Kleid DG (1975) Proc Natl Acad Sci USA 72:1184-1188 McKnight GL, Cardillo TS, Sherman F (1981) Cell 25:409-419 Messing J, Gronenborn B, Muller-Hill B, Hofschneider PH (1977) Proc Natl Acad Sci USA 74:3642-3646

C.F. Wright et al. : Expression Mutants in Yeast Montgomery DL, Leung DW, Smith M, Shalit P, Faye G, Hall BD (1980) Proc Natl Acad Sci USA 77:541-545 Montgomery DL, Boss JM, McAndrew SJ, Marr I, Walthall DA, Zitomer RS (1982) J Biol Chem 257:7756-7761 Ohtsubo H, Ohtsubo E (1978) Proc Natl Acad Sci USA 75: 615-619 Rave N, Crkvenjakov R, Boedtker H (1979) Nucleic Acid Res 6:3559-3567 Sanger F, Niclden S, Coulson AR (1977) Proc Natl Acad Sci USA 74:5463-5467 Sherman F, Stewart JW, Jackson M, Gilmore RA, Parker JH (1974) Genetics 77:255-284 Sherman F, Stewart JW, Margoliash E, Parker J, Campell W (1966) Proc Natl Acad Sci USA 55:1498-1504 Struhl K, Stinchcomb DT, Seherer S, Davis RW (1979) Proc Natl Acad Sci USA 76:1035-1039 Thomas PS (1980) Proc Natl Acad Sci USA 77:5201-5205 Williamson VM, Young ET, Ciriacy M (1981) Cell 23:605-614 Zitomer RS, Hall BD (1976) J Biol Chem 251:6320-6326 Zitomer RS, Montgomery DL, Nichols DL, Hall BD (1979) Proc Natl Acad Sci USA 76:3627-3631

Communicated by C. P. Hollenberg Received January 14, 1983