125

Mutation Research, 249 (1991) 125-133 © 1991 Elsevier Science Publishers B.V. 002%5107/91/$03.50 ADONIS 002751079100134L

MUT 04986

Induction of 3,8 transposition in response to elevated glucose-6-phosphate levels A n n e t t e T. L e e a n d A n t h o n y C e r a m i The Rockefeller University, Laboratory of Medical Biochemistry, New York, N Y 10021 (U.S.A.) (Received 17 October 1990) (Accepted 10 December 1990)

Keywords: Reducing sugar; y8 transposition; Escherichia coli

Summary The nonenzymatic glycosylation of nucleic acids in vitro by the reducing sugars, glucose or glucose-6phosphate, alters both physical and biological properties. Recent investigations have demonstrated that elevated intracellular levels of glucose-6-phosphate in glycolytic mutants of E. coli resulted in a concentration-associated increase in mutations of a target plasmid. The majority of the plasmid mutations were due to large ( > 1 kb) insertions or deletions. We describe here the further analysis of mutant plasmids isolated from bacteria grown under conditions which were conducive to the intracellular accumulation of glucose6-phosphate. We have found that a number of the insertional plasmid mutations were the result of the movement of the transposable element y8 from the host genome into the plasmid. The frequency of ,/8 transposition was also associated with the amount of glucose-6-phosphate accumulated in the bacterial cells. Furthermore, the presence of another transposable element, either T n 5 or T n l O in the host genome increased the rate of ,/8 transposition without affecting its own movement. The observed increase in ,/8 transposition suggests a novel mechanism of induction by reducing sugars which may be the result of D N A modifications by reducing sugars.

It has come to the attention of a number of investigators that reducing sugars such as glucose and glucose-6-phosphate can react with the amino groups of a number of biologically relevant molecules. The nonenzymatic reaction of reducing sugars with the amino groups of proteins was first described by the chemist L.C. Maillard (Maillard, 1912). This reaction is responsible for the golden

Correspondence: Dr. A.T. Lee, The Rockefeller University, Laboratory of Medical Biochemistry, 1230 York Ave., New York, NY 10021 (U.S.A.).

brown color of cooked food and the changes in taste and texture observed in foods following long term storage. Since its initial discovery, the nonenzymatic glycosylation of proteins by reducing sugars has stimulated much interest in both the food industry and the biological sciences. The Maillard reaction initially begins with the formation of a reversible Schiff base between the aldehyde of the reducing sugar and the amino group of a protein. Within a few weeks, the Schiff base reaches an equilibrium with a more stable, but still reversible Amadori product. The Amadori product itself can undergo a series of further de-

126

hydrations and rearrangements to form irreversibly an array of endproducts, collectively referred to as advanced glycosylation endproducts. These endproducts are characteristically fluorescent, yellow-brown in color, and are able to crosslink proteins inter- and intra-molecularly. The amount of advanced glycosylation endproducts formed is largely dependent upon sugar concentration and length of exposure of the protein to the sugar. The reaction of reducing sugars with long lived proteins such as lens crystallins (Monnier and Cerami, 1981; Tsilibary et al., 1987) and collagen (Stevens et al., 1978; Kohn et al., 1984; Monnier et al., 1984) has provided some potential explanations to account for some of the complications associated with diabetes mellitus and aging. Reducing sugars can react with long lived proteins to affect their physical and in some cases their biological properties. Several years ago, it was hypothesized that reducing sugars could also react with the amino groups of D N A bases in a manner analogous to the reaction observed with proteins and that this reaction of reducing sugars either alone or in a protein bound equivalent could contribute to some of the damage associated with the onset of aging. Bucala et al. (1984) demonstrated that in vitro incubation of the reducing sugar, glucose-6-phosphate (G-6-P), with single stranded DNA, double stranded D N A or nucleotides resulted in changes in absorbance and fluorescence spectra which were similar to the changes observed with proteins nonenzymatically glycosylated in vivo. Studies using single stranded fl phage D N A and double stranded pBR322 plasmid D N A confirmed that in addition to measurable physical changes, the in vitro glycosylation of nucleic acids by glucose or G-6-P resulted in alterations of some biological functions (Bucala et al., 1984, 1985). Incubations of fl D N A with glucose or G-6-P led to a time and sugar concentration dependent loss in transfection capacity (Bucala et al., 1984). The sugar modification of pBR322 D N A by G-6-P resulted in a comparable decrease in transformation efficiency. In addition to this decrease, the glycosylated plasmid D N A exhibited a higher mutation rate. The plasmid mutations were the result of assorted insertions or deletions (Bucala et al., 1985). It is interesting to note that an increase in

plasmid mutations was observed only when the glycosylated D N A was transformed into a repair proficient host which suggests that the observed mutations were not directly generated by the sugar modifications. Studies by Lee and Cerami (1987b) analyzed the effects of in vivo exposure of D N A to elevated G-6-P levels. By utilizing glycolytic mutants of E. coli which accumulated glucose-6-phosphate intracellularly when grown in the presence of glucose, it was demonstrated that in vivo exposure of a target plasmid to elevated G-6-P levels resulted in an increase in plasmid mutations which were associated with the amount of G-6-P accumulation. These studies showed that in comparison to the control strain K10, the DF40 strain which was deficient in phosphoglucose isomerase, accumulated a 20 fold increase in G-6-P and the DF2000 strain which was deficient in phosphoglucose isomerase and glucose-6-phosphate dehydrogenase accumulated a 30 fold increase, when grown in minimal medium containing glucose and gtuconate. Under these conditions, the mutation rate of the target plasmid increased 7 fold in the DF40 strain and 13 fold in the DF2000 strain when compared to the background rate observed in the K10 strain. Preliminary analysis of the plasmid mutations showed that a number of them were due to large ( > 1 kb) insertions and deletions. In this communication we report the further analysis of plasmid mutations due to insertions and demonstrate an increase in the transposition frequency of the transposable element, y6, from the host genome into the target plasmid in response to elevated intracellular G-6-P levels. We also report on the influence of elevated G-6-P levels on the transpositional rate of T n 5 and TnlO and the influence of these transposable elements on y6 transposition in our system. Materials and methods

Bacterial strains The following K12 E. coli strains were used:(i) K10 Hfr (tonA22, ornpF626, relA1, pit-lO, spoT1, T2R (E. coli Genetic Stock Center, New Haven, CT); CGSC4234); (ii) D F 4 0 H f r (tonA22,

ornpF627,

relA1,

pit-lO,

spoTl,

T_,R, pgi-2;

127 CGSC4861); (iii) DF2000 Hfr (tonA22, relA1, pit10, T~, pgi-2, zwfA2; CGSC4873); (iv) MC1060

F- ((Alacl-Y)74, galE15, galK16, ~-, relA1, rpsL150, spoT1, hsdR2; CGSC 6648); (v) SB4288 F- ((Alac-proB), recA, thi-1, relA, mal-24, spc12, supE-50, DE5); (vi) K38 Hfr (reel +, rel-1, tonA22, T~, phoA4, PO2A). E. coli strains K10, DF40 and DF2000 were separately transduced with ziel : : T n 5 or z b i : : T n l 0 (the generous gifts of Dr. Marjorie Russel) by P1 mediated transduction (Miller, 1972) then transformed with the plasmid pAM006. The transduced strains were tested for the presence of Tn5 or TnlO and will be referred to as K 1 0 / T n 5 , D F 4 0 / T n 5 , D F 2 0 0 0 / T n 5 , KIO/TnlO, DF40/TnlO and DF2OOO/TnlO, respectively. MC1060 was transduced with a P1 lysate from JC10240 Hfr (thr-300, ~-, recA56,

srl-3OO::TnlO, relA1, rpsE2300(SpcR), spoT1, ilv318, thi-1; CGSC6074) and colonies were selected for a l a c - / r e c A - / t e t R phenotype and will be referred to as MC1060 (recA56, srl-300::Tnl0).

Plasmids The plasmids pAM006 and pKM005 were a gift from Drs. P. Green (Rockefeller University, New York, NY) and M. Inouye (Rutgers University, New Brunswick, N J). pKM005 (9.9 kb) carries the gene conferring ampicillin resistance and inactive promoterless lacZ and lacY genes (Masui et al., 1983). pAM006 (10.1 kb) was derived by the insertion of the ompA promoter upstream of lacZ to activate lacZ and lacY transcription (Green, 1981). pAM006 was used as the test plasmid in our experiments, while pKM005 was used to confirm fidelity of our indicator strain MC1060 (recA56, srl-300::Tnl0) and tetrazolium lactose indicator plates. As internal controls, transformations using pAM006 and pKM005 were used routinely to verify our indicator strain and bacterial plates.

Media M63 minimal medium was prepared as described by Miller (1972) with the addition of 2% (wt/vol) glucose and gluconate in a 9:1 mass ratio, supplemented with ampicillin at 100 # g / m l . Tetrazolium lactose indicator plates were prepared as described by Miller (1972) and supplemented with ampicillin at 100 /~g/ml. LB broth (Gibco)

was prepared as directed and also supplemented with 100/~g/ml ampicillin.

Growth conditions The transformed strains were grown in 100 ml of M63 minimal medium containing glucose and gluconate for 24 h in a 37°C shaking water bath. Cultures were regularly checked for revertants in the mutants strains. Following growth each culture was harvested and plasmid D N A was isolated as described by Maniatis et al. (1982).

Detection of plasmid mutations The purified plasmid D N A from each strain was separately used to transform previously frozen competent MC1060 cells (recA56, s r l l 0 0 : : T n l 0 ) (Hanahan, 1985) and plated out on tetrazolium lactose indicator plates containing 100 /~g/ml ampicillin. The transformed colonies which were ampicillin resistant but unable to ferment lactose were isolated and rescreened for ampR/lac - phenotype. Phenotypically mutant colonies which were transformed with plasmid D N A from K 1 0 / T n 5 , D F 4 0 / T n 5 or D F 2 0 0 0 / T n 5 were rescreened on tetrazolium lactose indicator plates containing both ampicillin (100 # g / m l ) and kanamycin (20 # g / m l ) to assay for the transposition of T n 5 into the test plasmid. Plasmid D N A was isolated from all phenotypically mutant colonies and screened by Southern blot analysis for 338 insertion using a [3Zp]-kinased oligonucleotide ( T I T G A G G G C C A A T G G A A C G A ) which is homologous to both ends of ,[8 (Reed et al., 1979). Plasmid D N A which was isolated f r o m KIO/TnlO, DF40/TnlO, or DF2OOO/TnlO and resulted in phenotypically mutant colonies when transformed in the MC1060 (recA56, srll00:: T n l 0 ) indicator strain was analyzed for TnlO insertion by Southern Blot analysis using a [32p]_ kinased oligonucleotide ( C T G A T G A A T C C C CTAATGAT) which is homologous to both ends of TnlO (Kleckner, 1979).

Southern blot analysis Isolated plasmid D N A or genomic D N A isolated as described by Owen and Borman (1987) was digested with the indicated restriction enzymes, treated with RNAse A (50 /~g/ml) and electrophoresed through a 1% agarose gel contain-

128

ing 1 ~ g / m l ethidium bromide. The gel was denatured in transfer buffer (0.4 M N a O H / 0 . 6 M NaC1) for 30 min, then vacuum blotted (LKB, Vacugene, Vacuum Blotting System) in transfer buffer onto Nytran (Schleicher & Schuell). Upon completion of transfer, the membrane was neutralized in 0.5 M Tris-HC1, pH 7.0/1.0 M NaCI and baked at 65°C for 1 h. The blot was prehybridized and hybridized according to the manufacturer's instructions. Sequence analysis PstI fragments were isolated from a 1% agarose gel and electroeluted out of the gel (Maniatis et al., 1982). The fragments were separately cloned into pEMBL 1 8 ( + ) and transformed into JM109 cells. Small scale plasmid preparations were done on positive clones to determine orientation. Cells carrying the desired inserts were transfected with helper phage R408 and single stranded DNA was prepared (Russel et al., 1986). Sequence reactions were performed as instructed using Sequenase (US Biochemicals) and separated on a 6% acrylamide/ 8M urea sequencing gel. Sequence comparison was done using DNASIS v. 2.3. Results

Preliminary analysis of plasmid mutations of pAM006 ( l a c + ~ lac ) arising from elevated G-6-P levels in two mutants strains of E. coli demonstrated that a number of them were due to large ( > 1 kb) insertions or deletions. As an initial approach, we focused on the evaluation of insertional plasmid mutations. From restriction digest analysis with PstI, we found that a proportion of the mutated plasmids contained two characteristic fragments of 2.3 kb and 0.6 kb, in addition to other fragments (Fig. 1). Using these two fragments to probe Southern blots of bacterial genomic and wild type plasmid DNA, also digested with PstI (Fig. 2A), it was demonstrated that the 0.6 kb fragment was present in the genomes of the host strains, K10, DF40 and DF2000, but absent from the wild type plasmid (Fig. 2C). The 2.3 kb fragment also hybridized to the host strains and the wild type plasmid, but to a lesser extent indicating that the 2.3 kb fragment contains some plasmid sequences (Fig. 2B). The salient feature of

a

b

c

d

e

f

g

h

i

k

I

m

Kb

Kb

Fig. 1. Restriction digest analysis of mutated plasmid D N A with Pstl. Lanes b-1 are a representative sample of mutated plasmids; a, m, wild type plasmid D N A ; last lane, 1 kb D N A ladder markers (BRL). Plasmid D N A was isolated, digested with Pstl then electrophoresed through a 1.0% agarose gel containing 1 t t g / m l ethidium bromide.

these strains is the integration of the F episome into their respective genomes. Southern blot analysis of two other E. coli strains K38(Hfr) and SB4288(F-) confirmed that these fragments hybridized only to Hfr strains. Since the transposable element 3'8 is known to be on the F episome (Guyer, 1978), which was stably integrated into the Hfr strains (K10, DF40, DF2000 and K38), it was proposed that these fragments were the result of transposition of `/8 into the plasmid. The sequence of the 0.6 kb fragment was determined and found to be identical to the published sequence of the internal resolution site of the transposable element, 3'8 (Reed, 1981), confirming that transposition of y6 into the plasmid had occurred (sequence not shown). In order to quantitate the percentage of mutations due to 3'6 transposition, plasmid D N A from mutant colonies was analyzed on Southern blots using a [32p]-kinased oligonucleotide homologous to both ends of 3'6. Table 1 represents the number of mutant plasmids isolated from the K10, DF40 and DF2000 strains and the percentage of plasmid mutations which contained ,/6 transpositions observed. The plasmid mutation rate and the

129 frequency of "t8 t r a n s p o s i t i o n i n t o the m u t a t e d plasmids increased in response to the a c c u m u l a t e d intracellular level of G-6-P. The D F 2 0 0 0 strain

a

b

~' '

which a c c u m u l a t e d the highest level of G - 6 - P (0.864 + 0.011 /~mol G - 6 - P per 5 x 105 cells, Lee a n d Cerami, 1987b) exhibited the highest p l a s m i d

c

d

e

f

g

~i~ ii !?~"~ ~ ' ~ ....

h

i

~

Kb

~]1 ~ Kb

B

C a

b

c

d

e

f

g

h

a

b

c

d

e

f

g

i

h

I

-2.3

Kb

i

-0.6

Kb

Fig. 2. Southern blot analysis of genomic and plasmid DNA. (A) Genomic DNA was isolated from E. coli strains; (a) K10, (b) DF40, (c) DF2000, (d) K38, (e) SB4288 then digested with PstI and electrophoresed through a 1% agarose gel containing 1/~g/ml ethidium bromide. Lane f, wild type pAM006 DNA linearized with PstI, lanes g-i, plasmid DNA was isolated from mutant colonies then digested with PstI. The outer lane is a 1 kb molecular weight DNA ladder. (B) Southern blot analysis using [32P]-labeled2.3 kb PstI fragment derived from plasmid DNA in lane h. (C) Southern blot analysis using [32p]-Iabeled 0.6 kb PstI fragment derived from plasmid DNA in lane h.

130

mutation rate and the highest frequency of ./6 transposition into the plasmid. The DF40 strain which accumulated 20 times more G-6-P (0.553 + 0.072 /zmol G-6-P per 5 x 105 cells, ibid.) intracellularly than K10 (0.028 + 0.005 btmol G-6-P per 5 × 105 cells, ibid.) had lower plasmid mutation and y6 transposition frequencies than DF2000. Under these conditions there was no transposition of ],6 into any of the mutated plasmids from the K10 strain. Approximately 12% of the mutant plasmids from the DF40 strain were due to transposition of "/6 and almost 40% of the mutant plasmids from the DF2000 strain had a "/6 insertion. These results indicate that elevated intracellular levels of G-6-P not only led to an increase in plasmid mutation rate but also to an increased rate of "/6 transposition. These observations led to the hypothesis that elevated G-6-P levels and possibly D N A damage arising from the in vivo interaction of DNA with

G-6-P may act as a general induction signal of transposition. In order to investigate this possibility, two transposable elements, T n 5 and T n l 0 , which are unrelated to each other and y6 (Calos and Miller, 1980), were separately integrated into the genomes of K10, DF40 and DF2000 by P1 mediated transduction. These strains were subsequently grown under conditions which increased the rate of plasmid mutations and "/6 transposition in the DF40 and DF2000 strains. Table 1 demonstrates the profound effect of the presence of Tn5 in the DF2000 genome on the rate of plasmid mutation and frequency of ./6 transposition. The integration of Tn5 into the DF2000 strain increased the overall plasmid mutation rate 4 fold and the transposition of ./6 more than 7 fold. In the absence of Tn5, only 38% of the mutated plasmids had the 3,6 insertion, but in its presence, 69% of the mutated plasmids contained ./6 sequences. The plasmid mutation rate of DF40

TABLE 1 P L A S M I D M U T A T I O N S (Lac + ---, L a c - ) PER 103 T R A N S F O R M A N T S A N D P E R C E N T A G E O F -/8 T R A N S P O S I T I O N S K10

(% 3,8)

DF40

(% 3,8)

DF2000

(% 3,8)

(A) Plasmid mutations

0.40 _+0.62 0.0

(B) Plasmid mutations

(12)

0.89 ± 0.77

14.62 ___2.08

0.0

(38) 5.62 +_ 3.27

1.24 __+0.87

(0) y8 + Tn 5

14.62 ± 5.20

9.90 _+4.67

(o)

58.57 + 14.56 (35)

5.10 + 1.76

(69) 40.52 + 18.09

(c) Plasmid mutations

6.20 + 1.45

6.10 + 1.99 (32)

3,8 + T n l 0

2.00+0.0

13.47 _+ 4.70 (84)

5.10+ 1.96

(65) 8.71 + 3.47

E. coli strains K10, DF40 and DF2000 alone or containing either T n 5 or TnlO in their respective genome were grown in minimal medium containing glucose and gluconate for 24 h. Plasmid D N A was isolated from each strain and used to transform MC1060 (recA56, s r l l 0 0 : : T n l 0 ) competent cells. The transformants were grown on tetrazolium lactose indicator plates containing ampicillin (100 /xg/ml). Colonies which displayed a ampR/lac - phenotype were scored as a plasmid m u t a n t and rescreened for the transposition of 3'8, T n 5 or TnlO. (A) Plasmid D N A was isolated from all phenotypically m u t a n t colonies and screened by Southern blot analysis for the insertion of y8 by using a [32p]-kinased oligonucleotide as a probe. (B) Phenotypically m u t a n t colonies which were suspected of containing a T n 5 insertion were rescreened on tetrazolium lactose indicator plates containing arnpicillin (100/~g/rnl) and kanamycin (20 # g / m l ) . (C) Plasmid D N A s from phenotypically mutant colonies which were suspected of containing a TnlO insertion were analyzed by Southern blot analysis using a [32p]-kinased oligonucleotide as a probe.

131 also increased in the presence of T n 5 but to a lesser extent. The integration of Tn5 into the DF40 genome resulted in a 1.5 fold increase in plasmid mutations which was associated with a 4 fold increase in "/8 transposition. Transposition of "/8 into the target plasmid accounted for about 35% of the plasmid mutations, whereas in the absence of Tn5, only 12% of the plasmid mutations were due to "/8 insertion. A 2 fold increase in plasmid mutations was observed when T n 5 was present in the K10 genome, however none of these plasmid mutations were attributable to "/8 transposition. The transposition of T n 5 into the plasmid was not detected in any of the mutated plasmids isolated from the host strains. The presence of T n l O in the DF2000 genome had no effect on overall plasmid mutation rate. However, the percentage of plasmids containing "/8 insertions was raised to 65% which represented a 1.5 fold increase in "/8 transposition. The plasmid mutation rate of the DF40 strain decreased 1.6 fold, but the percentage of "/8 insertions increased from 12% to 84% in the presence of TnlO. When T n l O was present in the K10 genome, there was more than a 15 fold increase in the plasmid mutation rate, combined with the occurrence of "/8 transposition in 32% of the mutated plasmids isolated. As in the case of Tn5, no transposition of T n l O was observed in any of the mutant plasmids isolated; however, control studies demonstrated that both T n 5 and T n l O were capable of transposition into our target plasmid (data not shown). Discussion The increase in plasmid mutations (lac ÷ l a c - ) observed in response to elevated intracellular levels of G-6-P prompted the further investigation into the possible mechanisms responsible for this increase. Analysis of some of the mutant plasmids demonstrated the presence of two characteristic fragments of 2.3 kb and 0.6 kb following PstI digestion. Using these two fragments as probes, Southern blot analysis of genomic and wild type plasmid D N A showed them to be present only in the genomes of the test strains K10, DF40 and DF2000. One prominent feature of these strains is the integration of the F episome into their respective genomes. Southern blot and

sequence analysis of the 0.6 kb PstI fragment confirmed that T8 had transposed from the bacterial genome to the target plasmid (Fig. 2A, C). The plasmid mutation rate and the rate of ./8 transposition into the target plasmid increased in response to elevated intracellular levels of G-6-P. The DF2000 strain which accumulated the highest level of G-6-P (30 fold over control sample), when grown in minimal medium containing glucose, also exhibited the highest plasmid mutation rate and the highest frequency of T8 transposition. The DF40 strain which had a 20 fold increase in intracellular G-6-P levels had a lower plasmid mutation and T8 transposition rate than DF2000. Under these conditions there was no transposition of T8 into any of the background plasmid mutations isolated from the K10 control strain. The finding that the reducing sugar G-6-P, or possible D N A damage resulting from exposure to reducing sugars, could serve as an inducing signal for the transposition of TS, led to further investigations as to whether these same factors could induce the transposition of the unrelated transposable elements, T n 5 and TnlO. If the intracellular elevation of G-6-P levels or resulting D N A damage were general signals for the induction of transposition, then an increase in transposition of all three transposons would be observed in the target plasmid with exposure to increasing G-6-P levels. The presence of T n 5 in the genomes of DF40 and DF2000 increased both plasmid mutation rate and the transposition frequency of "/8. A dramatic increase in plasmid mutation rate and T8 transposition was observed in the DF2000 strain (see Table 1). Increases were also observed in the DF40 strain, but to a lower extent as the result of the integration of T n 5 into the respective genomes. No change in plasmid mutation rate was observed in the K10 strain. The integration of T n l O into the genomes of K10, DF40 and DF2000 strains affected the plasmid mutation rate and the transposition rate of ./8 in a similar manner, although some differences were evident (Table 1). A pronounced effect of T n l O integration was observed in the K10 strain, where there was an increase in plasmid mutations and the occurrence of ./8 transposition.

132 Although the plasmid mutation rate of the DF40 strain decreased with the presence of T n l 0 , the percentage of plasmid mutations due to T8 transposition was increased to 84% of the total plasmid mutations. The percentage of 3,8 transpositions into the target plasmid was increased to 65% in the DF2000 strain even though there was no change in the overall rate of plasmid mutations. The effects of T n l O integration appear to influence the plasmid mutation rate and "t8 transposition in a different manner than Tn5. As observed with the genornic integration Tn5, there is an increase in the transposition rate of ,/3 in the presence of TnlO. These increases in T8 transposition are associated with the relative levels of intracellular G-6-P found in each of the respective strains. However, the proportion of plasmid mutations due to T8 transposition in the K10 and DF40 strains is different when T n l O is present in the genome as compared to when Tn5 is present. In the case of T n l O integration into the DF2000 genome, the percentage of plasmid mutations due to T8 transposition remains comparable to that observed when T n 5 is present. It is surprising that the presence of either Tn5 or T n l O influences the frequency of T8 transposition without transposing itself. Control studies have demonstrated that T n 5 and T n l O were able to transpose into the target plasmid, although none were observed in the plasmid mutants isolated from the experimental samples. Our observations present a number of intriguing questions pertaining to the influence of G-6-P a n d / o r its modification of DNA in activating ,/3 transposition. It is possible that DNA modification or D N A damage by G-6-P directly induces ,/3 transposition. Bucala et al. (1985) reported the movement of the insertional element IS1 into the target plasmid, pBR322, following modification by G-6-P in vitro. Preliminary studies transforming K10, DF40 or DF2000 with plasmid DNA which had been glycosylated in vitro by an adduct of G-6-P and lysine, followed by growth under conditions which did not result in G-6-P accumulation, also led to an increase in plasmid mutations and transposition of T8 from the host genome into the plasmid (Lee and Cerami, 1987a and unpublished data). It appears the frequency of -/3 transposition is related to the level of G-6-P

accumulation and also possibly D N A modification by G-6-P; however, the presence of an additional transposable element, T n 5 or TnlO, synergistically increases the frequency of 3,8 transposition. These differences in the rate of T3 transposition associated with elevated G-6-P levels in the presence of T n 5 or T n l O may be related to the differences each of the respective elements transposes. The mechanism behind this apparent synergism is unknown but is under investigation. Signals inducing the transposition of transposable elements have not yet been elucidated, although an increase in transposition induced by chemical stimulation has been observed for the transposable elements, T n 9 1 7 (Tomich et al., 1980), Tn501 (Arthur et al., 1981) and T n 9 (Datta et al., 1983). Although it does not appear that elevated G-6-P level is a universal induction signal for transposition, it is evident that it can positively affect T8 transposition in our system. Recent investigations have demonstrated a causative role for mammalian transposable elements and insertional sequences in somatic mutations which were responsible for activation of cellular oncogenes (Rechavi et al., 1982; Cohen et al., 1983; Morse et al., 1988). Future investigations will be necessary to determine whether D N A modifications resulting from the time dependent reaction with reducing sugars can influence the transposition frequency of mammalian transposable elements. Experiments along these lines will lead to new insights into the mechanism behind the potential of normal cellular metabolites (e.g., G-6-P) to induce DNA rearrangements, influence the genetic composition in eukaryotic systems, and possibly explain the age dependent increase in the rate of cancer (Bennington, 1986).

Acknowledgements We would like to thank Dr. P.J. Green for pAM006 and Dr. M. Russel for zie-l::Tn5, zbi::Tnl0, and P1 lysate of JC10240. This research was supported under a N S F graduate fellowship awarded to A.T. Lee, a N I H (DK19655) grant to A. Cerami, and the Brookdale Foundation.

133

References Arthur, A., M. Burke and D. Sherratt (1981) Inter-replicon transposition of the Hg transposon Tn501 is inducible by Hg 2+ ions and proceeds through cointegrate intermediates, Soc. Gen. Microbiol. Quart., 96-102. Bennington, J.L. (1986) Cancer and aging - pathology, Front. Radiat. Ther. Oncol., 20, 45-51. Bucala, R., P. Model and A. Cerami (1984) Modification of DNA by reducing sugars: a possible mechanism for nucleic acid aging and age related dysfunctions in gene expression, Proc. Natl. Acad. SCi. (U.S.A.), 81, 105-109. Bucala, R., P. Model, M. Russel and A. Cerami (1985) Modification of DNA by glucose-6-phosphate induces DNA rearrangements in an Escherichia coil plasmid, Proc. Natl. Acad. Sci. (U.S.A.), 82, 8439-8442. Calos, M.P., and J.H. Miller (1980) Transposable elements, Cell, 20, 579-595. Cohen, J.B., T. Unger, G. Rechavi, E. Canaani and D. Givol (1983) Rearrangement of the oncogene c-mos in mouse myeioma NS1 and hybridomas, Nature (London), 306, 797-799. Datta, A.R., B.W. Randolph and J.L. Rosner (1983) Detection of chemicals that stimulate Tn9 transposition in Escherichia coil K12, Mol. Gen. Genet., 189, 245-250. Green, P.J. (1981) Ph.D. Thesis, State University of New York at Stony Brook, Stony Brook, NY. Guyer, M.S. (1978) The -f8 sequence of F is an insertion sequence, J. Mol. Biol., 126, 347-365. Hanahan, D. (1985) Techniques for transformation of E. coli, in: D.M. Glover (Ed.), DNA Cloning, A Practical Approach, Vol. 1, IRL Press, Washington, DC, pp. 109-135. Kleckner, N. (1979) DNA sequence analysis of T n l O insertions: origin and role of 9 bp flanking repetitions during T n l O translocation, Ceil, 16, 711-720. Kohn, R.R., A. Cerami and V.M. Monnier (1984) Collagen aging in vitro by nonenzymatic glycosylation and browning, Diabetes, 33, 57-59. Lee, A.T., and A. Cerami (1987a) The formation of reactive intermediates capable of rapidly reacting with DNA, Mutation Res., 179, 151-159. Lee, A.T., and A. Cerami (1987b) Elevated glucose 6-phosphate levels are associated with plasmid mutations in vivo, Proc. Natl. Acad. Sci. (U.S.A.), 84, 8311-8314. Maillard, L.C. (1912) Action des aeides amin~ sur les sucres; formation des m61anoidines par voie m&hodique, C.R. Seances Acad. SCi., 154, 66-68.

Maniatis, T., E.F. Fritsch and J. Sambrook (1982) Molecular Cloning; A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Masui, Y., J. Coleman and M. Inouye (1983) Experimental Manipulation of Gene Expression, Academic Press, New York. Miller, J.H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Monnier, V.M., and A. Cerami (1981) Nonenzymatic browning in vivo: possible process for aging of long lived proteins, Science, 211,491-493. Mormier, V.M., R.R. Kohn and A. Cerami (1984) Accelerated age-related browning of human collagen in diabetes mellitus, Proe. Natl. Acad. Sci. (U.S.A.), 81, 583-587. Morse, B., G. Rotherg, V.J. South, J.M. Spandorfer and S.M. Astrin (1988) Insertional mutagenesis of the myc locus by a LINE-1 sequence in a human breast carcinoma, Nature (London), 333, 87-90. Owen, R.J., and Borman, P. (1987) A rapid biochemical method for purifying high molecular weight bacterial chromosomal DNA for restriction enzyme analysis, Nucleic Acids Res., 15, 3631. Rechavi, G., D. Givol and E. Canaarti (1982) Activation of a cellular oncogene by DNA rearrangement: possible involvement of an IS-like element, Nature (London), 300, 607-611. Reed, R.R., R. Young, J.A. Steitz, N. Grindley and M. Guyer (1979) Transposition of the Escherichia coli insertion element ,t8 generates a five base pair repeat, Proc. Natl. Acad. Sci. (U.S.A.), 76, 4882-4886. Reed, R.R. (1981) Resolution of cointegrates between "tO and Tn3 defines the recombinational site, Proc. Natl. Acad. SCi. (U.S.A.), 76, 3428-3432 Russei, M., S. Kidd and M.R. Kelley (1986) An improved filamentous helper phage for generating single stranded plasmid DNA, Gene, 45, 333-338. Stevens, V.J., C.A. Rouzer, V.M. Mormier and A. Cerami (1978) Diabetic cataract formation: potential role of glycosylation of lens crystallins, Proc. Natl. Acad. SCi. (U.S.A.), 75, 2918-2922. Tomich, P.K., F.Y. An and D.B. Uwell (1980) Properties of erythromycin inducible transposon Tn917 in Streptococus faecalis, J. Bacteriol., 141, 1366-1374. Tsilibary, E.C., A.S. Charonis and L.T. Furcht (1987) Changes of the main noncollagenous NC1 domain of type IV collagen after in vitro nonenzymatic giucosylation, Diabetes, 36, 85A.