Transcription increases the deletion frequency of long CTG[middot]CAG triplet repeats from plasmids in Escherichia coli
Transcription increases the deletion frequency of long CTG[middot]CAG triplet repeats from plasmids in Escherichia coli Richard P. Bowater+, Adam Jaworskiw, Jacquelynn E. Larson, Pawel Parniewski and Robert D. Wells*
Albert B.Alkek Institute of Biosciences and Technology, Texas A&M University, Center for Genome Research, Department of Biochemistry and Biophysics, Texas Medical Center, 2121 West Holcombe Boulevard, Houston, TX 77030, USA
Received March 20, 1997;Revised and Accepted May 22, 1997
ABSTRACT
Induction of transcription into long CTG[middot]CAG repeats contained on plasmids in Escherichia coli is showntoincrease the frequency of deletions within the repeat sequences. This elevated genetic instability was detected because active transcription into the triplet repeat influenced the growth transitions of the host cell, allowing advantageous growth for cells harboring plasmids with deleted repeat sequences. The variety of deletion products observed in separate cultures suggests that transcription altered the metabolism of the DNA in a manner that produced random length changes in the repeat sequence. For cultures containing plasmids without active transcription into the triplet repeat, or those maintained in exponential growth, deletions occurred within the repeat at a lower frequency (5-20-fold lower). In these incubations the extent of deletions was proportional to the number of cell divisions and many repeat lengths were observed within each culture, suggesting that the decrease in average repeat length at long incubation times was due to multiple small deletions. These observations show that deletions within long CTG[middot]CAG repeats contained on plasmids in E.coli occur via more than one pathway and their level of genetic instability is altered by the enzymatic processes occurring upon the DNA.
INTRODUCTION
Simple repetitive DNA sequences, which are widespread throughout natural genomes, are polymorphic in length due to intrinsic genetic instabilities (1 -3 ). The recent surge of interest in the factors influencing such instabilities is due to the association of simple repetitive DNA sequences with a variety of human diseases (4 ,5 ). For example, hereditary non-polyposis colon cancer cell lines have a high frequency of length changes in specific dinucleotide repeats (6 ,7 ) and unusual mutation events, involving the expansion of triplet repeat sequences (TRS), are associated with a number of neurogenetic disorders (for recent reviews see 8 -12 ). The propensity for TRS to expand provides an explanation for the non-Mendelian inheritance patterns (termed anticipation) of their associated diseases; with progression through a pedigree, the TRS become longer and the disease has an earlier age of onset and an increased severity.
Some features of the events leading to certain diseases associated with expansion of TRS are known (11 ,12 ). The expansion events fall into two categories, termed Types 1 and 2 (12 ). In Type 1 expansions, the TRS is always CTG[middot]CAG and is in a coding segment of the gene. In this case expansion events occur in small steps and the net effect is to produce proteins with longer tracts of polyglutamines. It is likely that these mutations lead to disease proteins with a gain of function that is particularly deleterious in neurons. Type 2 expansion events are associated with multi-system disorders and have a number of differences; various types of sequences are found and the TRS is not located within a coding region of the gene. All of the Type 2 expansion events are extremely large (hundreds of copies of the repeat for disease alleles).
Although progress has been made in understanding the pathologies of diseases associated with genetic instability of simple repetitive DNA sequences, the mechanism of expansion of TRS remains unclear. A number of studies have shown that sequences which undergo expansion can adopt unusual structures (reviewed in 13 ) which may cause aberrant metabolism of the DNA. To attempt to elucidate these mechanisms, we have initiated in vivo studies in Escherichia coli. Maintenance of long TRS in bacterial systems is difficult since they are deleted readily to sequences of heterogeneous length when contained on plasmids (14 ,15 ). Despite these problems, expanded products of some TRS have been obtained in E.coli (14 ,16 ), although the relationship of these observations to eukaryotic systems remains to be determined.
Herein we show that the deletion of long CTG[middot]CAG tracts from plasmids in E.coli is increased upon active transcription into the repeat sequence. An elevated frequency of deletions is observed because cells harboring plasmids with deleted repeat sequences have a growth advantage as the cultures pass through stationary phase (17 ). In the absence of transcription, a lower rate of deletions is observed and their extent was dependent on the number of cell division events. Hence, at least two pathways leading to deletions of TRS exist in E.coli.
MATERIALS AND METHODS
Plasmids and bacterial strains
Plasmids used in these experiments contained repeating CTG[middot]CAG inserts, which may also be designated TGC[middot]GCA or GCT[middot]AGC, have been described previously (14 ,17 ). The orientation of this TRS with respect to the direction of replication within these plasmids has been defined (14 ) as follows: plasmids containing the CTG sequence as the leading strand template are designated orientation I; plasmids containing the CTG sequence as the lagging strand template are designated orientation II. Figure 1 shows maps of pRW3247 and pRW3268 with the restriction sites relevant to this study. Both plasmids contain the sequence (GCT)27ACT(GCT)40ACT(GCT)106 [referred to as (CTG)175[middot](CAG)175 for convenience] in orientation II; note that pRW3247 has an intact promoter for lacZ', but pRW3268 does not. Equivalent plasmids with the same TRS in orientation I are pRW3248 (intact lacZ' promoter) and pRW3269 (deleted lacZ' promoter). In some experiments plasmid pIQ-kan (18 ) was used to provide the lacIQ repressor inside the cell.
Large amounts of plasmids were prepared by alkaline lysis of a 1 l culture (LB + antibiotic) grown to stationary phase and were purified by CsCl/ethidium bromide ultracentrifugation overnight (19 ). Purification of plasmids from 10 ml cultures was performed using the standard alkaline lysis miniprep procedure (19 ).
All cloning procedures and experiments were conducted in E.coli HB101 as described previously (17 ). After transformations, bacteria were allowed to recover at 37oC in SOC medium (2% Bacto-tryptone, 0.5% yeast extract, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, 10 mM MgSO4, 20 mM glucose, pH 7.0) for 30-45 min and then grown up in liquid culture with antibiotic selection; ampicillin and kanamycin were included at final concentrations of 100 and 50 [mu]g/ml respectively. When required, IPTG was included in the growth medium at a final concentration of 1 mM. To check the efficiency of each electroporation, 20 [mu]l of the transformation mixture were grown overnight at 37oC on LB-agar plates containing antibiotic.
Conditions of bacterial growth
The genetic stability of plasmids containing TRS was analyzed under a variety of conditions. In the basic procedure, E.coli HB101 was transformed with the appropriate plasmid and an aliquot was inoculated into glass tubes (150 mm length * 17 mm diameter) containing 10 ml LB with antibiotic. For incubations containing IPTG, the transformed cells were first grown for 2 h without IPTG to allow cell recovery in the absence of transcription from the lacZ' promoter and this culture was then diluted into fresh medium with or without IPTG.
Incubation of the liquid cultures was continued for the required time at 37oC and a shaking rate of 250 r.p.m. Growth of the cultures was monitored by optical density readings at 600 nm. The bacteria were subcultured into fresh liquid medium as appropriate: inoculations from stationary phase cultures were performed after 24 h (OD600 >2) with a dilution factor of 107; inoculations from logarithmic phase cultures were performed after 12 h (OD600 0.3-0.8) with a dilution factor of 4 * 106. The cells from each culture were harvested and stored at -80oC until the end of the experiment and then all DNA minipreps were performed at the same time.
Analysis of DNA
Restriction digests were performed following the manufacturer's instructions and the samples were radiolabeled at the EcoRI site using E.coli DNA polymerase I Klenow fragment incubated with [[alpha]-32P]dATP (19 ). The samples were analyzed by electrophoresis through 5% polyacrylamide gels in TBE buffer and the radioactive fragments were observed by autoradiography or with a phosphorimager (400S, Molecular Dynamics). Quantitation of the gels was performed directly upon the phosphorimage.
RESULTS
Transcription into (CTG)175[middot](CAG)175 increases its frequency of deletions
Long tracts of CTG[middot]CAG repeats are prone to deletion over extended periods of growth in E.coli. Previous investigations (14 ,17 ) showed that (CTG)175[middot](CAG)175 was more stable when the CTG strand was the leading strand template for replication (orientation I) compared with when it was the lagging strand template (orientation II). For all plasmids containing long CTG[middot]CAG tracts, deletions were observed within the TRS. However, the frequency of deletions varied widely between different vectors. Diverse aspects of DNA metabolism (e.g. recombination, repair, replication, transcription) were proposed to explain such variations, but the relative importance of each was not known. Thus the present experiments were undertaken to identify the factors influencing the frequencies of deletions of long CTG[middot]CAG repeats from plasmids in E.coli.
In an earlier study (14 ) the following pairs of plasmids were produced: pRW3247 (orientation II) and pRW3248 (orientation I); pRW3268 (orientation II) and pRW3269 (orientation I). These plasmids all contain (GCT)27ACT(GCT)40ACT(GCT)106 [referred to as(CTG)175[middot](CAG)175], but growth of E.coli harboring these plasmids produced different frequencies of deletions within each of the TRS (14 ). An obvious difference in the two sets of plasmids is that pRW3247 and pRW3248 have an intact promoter for lacZ', whereas pRW3268 and pRW3269 do not; thus the TRS is located inside a transcription unit for pRW3247 and pRW3248, but not for pRW3268 and pRW3269 (Fig. 1 ).
Transcription influences the genetic stability of triplet repeats in cultures that pass through stationary phase
Growth phase influences deletion frequency only when the TRS is located within an active transcription unit
Thus far, we have shown that active transcription into the TRS of plasmids in E.coli HB101 produced an increased frequency of deletions in repeat sequences from cultures that passed through stationary phase compared with those maintained in exponential growth. Since pRW3268 and pRW3269 do not have a promoter for lacZ', the frequency of deletions from the TRS in these plasmids should not be influenced by growth phase. Therefore, experiments were performed with pRW3268 and pRW3269 to test this hypothesis. Since transcription from the lacZ' promoter was not possible for these plasmids, in these experiments pIQ-kan was not co-transformed into the bacteria and IPTG was not included in the growth medium.
Active transcription into TRS produces a growth advantage for cells with deleted products after passage of the culture through stationary phase
We wished to determine whether the advantageous propagation of cells harboring deleted repeats was due to effects of transcription at a specific growth phase of the culture. Thus, an experiment was performed in which the plasmid harboring an intact lacZ' promoter with the TRS in orientation II (pRW3247) was transformed into E.coli HB101 containing pIQ-kan and at each subculturing the cells were diluted into medium with or without IPTG.
The strategy and results of the experiment are shown in Figure 5 . The proportion of plasmid containing full-length TRS was quantitated from restriction digests analyzed on polyacrylamide gels (data not shown). The results showed that the genetic stability of the TRS was determined by the state of the parent culture regarding transcription into the TRS. If the parent culture had transcription into the TRS then, after passage through stationary phase, its daughter cultures had increased amounts of deletions with and without IPTG. Conversely, if there was no active transcription into the TRS in the parent culture, the daughter cultures had relatively little change in their frequency of deleted products, even when active transcription into their TRS was induced by the presence of IPTG. In a control experiment the cultures were washed and incubated with medium with or without IPTG immediately prior to subculturing; as expected, the presence or absence of IPTG had no effect at this stage (data not shown).
Thus active transcription into (CTG)175[middot](CAG)175 contained on plasmids influences the transitions between stationary and exponential growth. This produces advantageous growth for cells harboring plasmids with deleted TRS and increases the observed amount of deletions in the TRS in cultures that pass through stationary phase.
DISCUSSION
The genetic stability of simple repeating sequences contained on plasmids in E.coli is influenced by many factors. Long CTG[middot]CAG repeats were shown to have a higher frequency of deletions upon active transcription into the TRS. This elevated frequency of deletions (5-20-fold increase for each growth cycle, depending on the experiment) was not proportional to the number of generations of the culture, but was produced upon passage of cultures through stationary phase. This observation explains earlier experiments which detected a growth advantage for cells harboring plasmids which contained deleted TRS (17 ). These results are unlikely to be due to an elevated copy number of plasmids harboring deletions because the size of the TRS has little effect on the copy number of these plasmids (data not shown). Although growth advantages were obtained only for plasmids with active transcription into the TRS, we are uncertain whether the destabilizing influence was transcription or some other event dependent upon transcription, such as translation. Plasmids without active transcription into the TRS also produced deletions within the repeat, but at the lower frequency of cultures maintained in exponential growth and the extent was proportional to the number of cell division events. Hence, deletions from long CTG[middot]CAG repeats on plasmids in E.coli occur via more than one pathway.
The molecular processes responsible for the effect of transcription on deletion frequency are uncertain from this in vivo study. However, examination of the patterns of deletions from different plasmids with the TRS in orientation II (CTG sequence as lagging strand template) reveals details of the mechanism by which transcription influences the stability of long TRS. For example, in conditions providing a growth advantage for cells with deleted products, passage through stationary phase of cultures which harbor plasmids with transcription into the TRS (pRW3247 co-transformed with pIQ-kan and in the presence of IPTG), the pattern of deletions varied in different experiments and the frequency of deletions varied by up to 2-fold (see for example Figs 2 and 3 ). However, each culture contained plasmids with relatively few different lengths of TRS. Comparison of separate inoculations showed that a wide variety of lengths of TRS were produced (compare Figs 2 and 3 ); frequently the deletions produced only plasmids containing small repeats.
ACKNOWLEDGEMENTS
These studies were supported by grants from the National Institutes of Health (GM52982) and the Robert A.Welch Foundation. We thank D.M.J.Lilley and D.Chen for provision of the plasmid pIQ-kan and S.Kang, K.Ohshima, E.Lawson, R.Iyer and L.Bowater for helpful discussions.
REFERENCES
1 Charlesworth,B., Sniegowski,P. and Stephan,W. (1994) Nature, 371, 215-220.MEDLINE Abstract
2 Richards,R.I. and Sutherland,G.R. (1992) Cell, 70, 709-712.MEDLINE Abstract
38 JaworskiA., Blaho,J.A., Larson,J.E., Shimizu,M. and Wells,R.D. (1989) J. Mol. Biol., 207, 513-526.MEDLINE Abstract
*To whom correspondence should be addressed. Tel: +1 713 677 7651; Fax: +1 713 677 7689; Email: rwells@ibt.tamu.edu
Present addresses: +Imperial Cancer Research Fund, Clare Hall Laboratories, South Mimms, Hertfordshire EN6 3LD, UK and [sect]Department of Genetics of Microorganisms, Institute of Microbiology and Immunology, University of Lodz, 12/16 Banacha St., 90-237 Lodz and Center for Microbiology and Virology, Polish Academy of Sciences, 106 Lodowa St., 93-232 Lodz, Poland
