Nucleotide Sequence and Characterization of the pyrF Operon of Escherichia coli K12*

The pyrF gene of Escherichia coli K12, which en- codes the pyrimidine biosynthetic enzyme orotidine-5’-monophosphate (OMP) decarboxylase, is part of an operon that includes a downstream gene designated orfF. The orfF gene product is a small polypeptide of unknown function. The nucleotide sequence of a 1549-base pair chromosomal fragment containing this op- eron was determined. An open reading frame capable of encoding the 27-kDa OMP decarboxylase subunit was identified and shown to be the pyrF structural gene by purifying and characterizing OMP decarbox- ylase. The subunit molecular weight (Mr amino-terminal amino acid sequence, and amino acid composition of the polypeptide predicted from the nu- cleotide sequence are in excellent agreement with those properties determined for the purified enzyme. The orfF structural gene was tentatively identified and apparently encodes an 11,396-dalton polypeptide. The orfF translational initiation codon overlaps the pyrF termination codon, which may indicate translational coupling in the expression of these genes. The pyrF promoter was mapped by primer extension of in vivo transcripts. The primary transcriptional initiation site is 51 base pairs upstream of the pyrF structural gene. The level of pyrF transcription and OMP decarboxyl- ase synthesis was found to be coordinately derepressed by pyrimidine limitation, indicating that regulation of pyrF gene expression occurs at the transcriptional level. Inspection of the nucleotide was calculated from the recoveries of 10 relatively acid-stable amino acids (i.e. Asp, Thr, GIu, GIy, Ala, Leu, Phe, Lys, His, and Arg). The nanomoles of each of these amino acids c/ml was divided by the number of residues of that amino acid/subunit predicted from the DNA sequence. The conversion factor was the average of these quotients.

The pyrF gene of Escherichia coli K12, which encodes the pyrimidine biosynthetic enzyme orotidine-5'-monophosphate (OMP) decarboxylase, is part of an operon that includes a downstream gene designated orfF. The orfF gene product is a small polypeptide of unknown function. The nucleotide sequence of a 1549base pair chromosomal fragment containing this operon was determined. An open reading frame capable of encoding the 27-kDa OMP decarboxylase subunit was identified and shown to be the p y r F structural gene by purifying and characterizing OMP decarboxylase. The subunit molecular weight (Mr = 26,350), amino-terminal amino acid sequence, and amino acid composition of the polypeptide predicted from the nucleotide sequence are in excellent agreement with those properties determined for the purified enzyme. The orfF structural gene was tentatively identified and apparently encodes an 11,396-dalton polypeptide. The orfF translational initiation codon overlaps the pyrF termination codon, which may indicate translational coupling in the expression of these genes. The pyrF promoter was mapped by primer extension of in vivo transcripts. The primary transcriptional initiation site is 51 base pairs upstream of the pyrF structural gene. The level of pyrF transcription and OMP decarboxylase synthesis was found to be coordinately derepressed by pyrimidine limitation, indicating that regulation of p y r F gene expression occurs at the transcriptional level. Inspection of the nucleotide sequence indicates that pyrF gene expression is not regulated by an attenuation control mechanism similar to that described for thepyrBI operon or p y r E gene. Finally, we compared the amino acid sequences of the OMP decarboxylases from E. coli, Saccharomyces cerevisiae, Neurospora crassa, and Ehrlich ascites cells to identify conserved regions.
The pyrF gene of Escherichia coli K12 encodes the pyrimidine biosynthetic enzyme orotidine-5'-monophosphate (OMP)' decarboxylase (EC 4.1.1.23). This enzyme catalyzes * This work was supported by National Institutes of Health Grant GM 29466. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted 502768.

to the GenBankTM/EMBL Data Bank with accession number(s)
$ To whom correspondence should be addressed Dept. of Microbiology, University of Alabama a t Birmingham, University Station, Birmingham, AL 35294.
' The abbreviations used are: OMP, orotidine 5'-monophosphate; bp, base pairs; kb, kilobase pairs. the sixth and final step in the de novo synthesis of UMP, the precursor of all pyrimidine nucleotides. The enzyme appears to be catalytically active as a dimer of identical 27-kDa subunits (I).' The pyrF gene is not closely linked on the E. coli chromosome to any of the genes or operons that encode the other five pyrimidine biosynthetic enzymes (i.e. carAB, pyrBI, and pyrC-E). Recent studies have shown, however, that the pyrF gene is part of an operon in which there is a downstream gene (designated orfF in this paper) that encodes an approximately 13-15-kDa polypeptide of unknown function (2,3). The expression of the pyrF gene is negatively regulated over a t least a 6-fold range by pyrimidine availability. The regulatory effector appears to be a uridine nucleotide, but its exact identity is unknown (4). There is nothing known about the mechanism in which the uridine nucleotide functions, although there has been one report of a possible transacting regulatory factor (5).
This study represents an initial step in the elucidation of the molecular mechanism controlling pyrF gene expression. We report the nucleotide sequence of the pyrF operon, the identification of the structural genes, the mapping of the pyrF promoter, and the effect of pyrimidine availability on pyrF transcription in uivo. The results indicate that the mechanism regulatingpyrF gene expression is unlike the previously characterized UTP-sensitive attenuation control mechanisms of the pyrBI operon (6-11) and the pyrE gene (12, 13). Also included in this paper is a brief comparison of the amino acid sequences of the OMP decarboxylases of E. coli, Saccharomyces cereuisiae, Neurospora crassa, and mouse-derived Ehrlich ascites cells.

MATERIALS AND METHODS
Bacterial Strains and Plasmids-The strains used in this study are E. coliK12. Strain CLT43 (F -A(argF-lac)U169rpsLlSOthiAI relAl deoC7 ptsF25 flbB5301 rbsR car-94 recA56 srl-30O:Tn10) was constructed in this laboratory as previously described (10). Strain SK4766 (pyrFxTn5 Kanr/pDK26 bla+pyrF+) was generously provided by William Donovan and Sydney Kushner (University of Georgia). This strain carries the multicopy plasmid pDK26 (4.8 kb), which was constructed by ligating a 2.2-kb fragment of E. coli K12 chromosomal DNA that includes the pyrF gene to the 2.6-kb PuuII fragment of plasmid pBR325 (3). The 2.6-kb PuuII fragment includes the origin of replication and the bla gene of plasmid pBR325 (14). Strain CLT43/pDK26 was constructed by transforming strain CLT43 with plasmid pDK26. D N A Preparations and Oligodeozynucleotide Synthesis-Plasmid DNA, bacteriophage Ml3mplO DNA, and DNA restriction fragments were prepared as previously described (10). Oligodeoxynucleotides were synthesized using an Applied Biosystems Model 380A DNA synthesizer and were purified by gel electrophoresis (10). DNA was * A smaller subunit molecular mass has also been reported (2); however, it was subsequently discovered that the gene encoding this polypeptide contained a deletion (K. F. Jensen, personal communication). 5'-32P-end-labeled as previously described (15).
DNA Sequence Analysis-DNA sequences were determined by the method of Maxam and Gilbert (15) and by the dideoxynucleotide chain termination method of Sanger et al. (16) using recombinant Ml3mplO templates and synthetic oligodeoxynucleotide primers.
Media and Culture Methods-Cells used for the isolation of RNA and for OMP decarboxylase assays were grown in N-C-medium (17) supplemented with 10 mM NHICl, 0.4% glucose, 0.05 mM thiamine hydrochloride, 1 mM arginine, 25 pg of ampicillin/ml (with strain CLT43/pDK26 only), and either 1 mM uracil or 71 p~ UMP. Cultures (100 ml in a 500-ml flask) were grown with shaking at either 30 or 37 "C as indicated.
Isolation of Cellular RNA-Cellular RNA used for mapping pyrF transcripts was prepared from strain CLT43/pDK26 by a procedure similar to that described by Hagen and Young (18). A 20-ml sample of exponential phase culture (ASSO = 0.5) was rapidly added to 2 ml of 10 X RNA extraction buffer (0.5 M Tris-HC1 (pH 6.8), 20 mM Na2EDTA, 10% sodium dodecyl sulfate) in a flask held in a boiling water bath. After 2 min, the flask was removed from the water bath and allowed to cool to room temperature, and 2 ml of 2 M sodium acetate (pH 5.2) was added to the cell lysate. The lysate was extracted twice with an equal volume of water-saturated phenol and once with an equal volume of chloroform. The aqueous phase was dialyzed phosphate. The RNA was precipitated from the dialysate by adding overnight against 2 liters of sterile water (4 "C) to remove interfering one-tenth volume of 3 M sodium acetate (pH 5.2) plus 2.5 volumes of 95% ethanol and incubating for 30 min at -70 "C. The RNA was collected by centrifugation (4 "C), dissolved in 0.3 ml of water, transferred to a 1.5-ml Eppendorf tube, and precipitated as described above. The RNA was collected by centrifugation, washed once with 0.5 ml of 70% ethanol (-20 "C), dried in UQCUO, and dissolved in 0.2 ml of DNase I buffer (10 mM Tris-HC1 (pH 8.0), 5 mM MgClz, 0.13 mM CaC12). Contaminating DNA was removed from the sample by adding RNase-free DNase I (40 pg/ml, Boehringer Mannheim) and incubating at 37 "C for 15 min. The RNA sample was then extracted once each with an equal volume of phenol and chloroform and precipitated as described above. The RNA was collected, washed, and dried as described above. The RNA was dissolved in 0.1 ml of water and stored at -20 "C. Each step of the RNA isolation procedure was performed quantitatively to permit a comparison of pyrF transcript levels in cells grown under different conditions.
Primer Extension Mapping-The 5' termini of pyrF transcripts were mapped by primer extension essentially as described (19). Cellular RNA was mixed with an appropriate 5'-32P-end-labeled DNA primer (see text), and the nucleic acid was dried in UQCUO. This sample was dissolved in 25 pl of hybridization buffer (20 mM Tris-HC1 (pH 8.0), 0.1 M NaCl, 0.1 mM NazEDTA), heated at 100 "C for 2 min, and then allowed to hybridize at 60 "C for 4 h. After cooling to room temperature, 25 pl of RT buffer (0.2 M Tris-HC1 (pH 8.0), 20 mM MgCl,, 0.1 M KC1, 20 mM ditbiothreitol, and 1 mM each of dATP, dGTP, dCTP, and dTTP) and 12.5 units of avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim) were added to the sample. The sample was incubated for 1 h at 37 "C. The primer extension products were precipitated by adding 2.5 volumes of 95% ethanol and incubating the samples in a dry ice/ethanol bath for 15 min. After centrifugation, the pelleted sample was washed with 70% ethanol (-20 "C) and dried in uacuo. The dried sample was dissolved in DNA sequencing dye and analyzed by polyacrylamide gel electrophoresis alongside appropriate DNA sequencing reactions (15).
Preparation of Cell-free Extracts-Cell-free extracts of strains CLT43/pDK26 and CLT43 were prepared to measure OMP decarboxylase levels. Cultures were grown to an A,,, of 0.5, and either a 40-ml (CLT43/pDK26) or an 80-ml (CLT43) sample was taken. Cells were collected by centrifugation (4 "C), washed with 0.1 M Tris-HC1 (pH 7.8), and stored for 1 day at -70 "C without loss of enzymatic activity. Cells were resuspended in 5 ml of 0.1 M Tris-HC1 (pH 7.8), 5 mM P-mercaptoethanol and disrupted by sonic oscillation at 0 "C. Cell debris was removed from the sonicate by centrifugation at 17,000 x g for 30 min at 4 "C. This extract was used to measure enzyme activity and protein concentration.
OMP Decarboxylase Assay-The enzyme assay is based on the decrease in absorption at 285 nm which occurs when OMP is converted to UMP by OMP decarboxylase (A6 = 2.25 X lo3 M" cm") (1). Reaction mixtures (1 ml) contained 64 mM Tris-HC1 (pH 7.8), 5 mM P-mercaptoethanol, 0.15 mM OMP (trisodium salt, Sigma), and an enzyme sample. Reactions were initiated by the addition of enzyme and were incubated at 23 "C. The rate of decrease in absorbance was measured with a Gilford Model 260 recording spectrophotometer.
One unit of enzyme activity is defined as that amount which catalyzes the formation of 1 pmol of UMP/min under the conditions described.
Protein Determinations-Protein concentrations were measured using the method of Bradford (20) with crystalline bovine serum albumin used as the standard.
Quuntitation of Plasmid DNA-Cells from triplicate 10-ml samples of culture (A650 = 0.5) were harvested by centrifugation (4 "C) and stored at -70 "C until analyzed. Plasmid DNA was extracted, and relative levels were measured by densitometric scanning of photographic negatives of ethidium bromide-stained agarose gels as previously described (10).
Purification of OMP Decarboxylase-Strain CLT43/pDK26 was grown in 2 liters of N-C-medium (17) supplemented with 10 mM NH,Cl, 0.4% glucose, 0.015 mM thiamine hydrochloride, 1 mM arginine, 0.24 mM UMP, and 25 pg/ml ampicillin. The culture was incubated overnight at 37 "C with shaking. Cells were harvested by centrifugation, washed with 64 mM Tris-HC1 (pH 7.8), and stored at -70 "C. Subsequent steps were performed at 0-4 "C. A 7.3-g sample of cell paste was thawed and resuspended in 73 ml of 64 mM Tris-HCl (pH 7.8), 5 mM p-mercaptoethanol (buffer A). The cells were broken by sonic oscillation, and cell debris was removed by centrifugation at 48,000 X g for 30 min. The centrifugation step was repeated to remove additional debris. The crude extract (76 ml) was divided into seven aliquots which were placed in 125-ml flasks. An equal volume of buffer A containing 20 mM UMP was added to each flask. The flasks were gently swirled in a 63 "C water bath for 7.5 min and then immediately chilled on ice. Denatured protein was removed from the samples by centrifugation at 17,000 X g for 15 min. The supernatants were pooled (136 ml). Solid ammonium sulfate was slowly added with gentle stirring to the heat-treated extract until 55% saturation was reached. After 1 h of equilibration with slow stirring, the suspension was centrifuged at 12,000 X g for 15 min. The supernatant was recovered and brought to 65% saturation with the addition of solid ammonium sulfate. The sample was equilibrated and centrifuged as described above. The 55-65% pellet was recovered and dissolved in buffer A. The final volume of the solution was 3.0 ml. This sample was applied to a Sephadex G-100 column (1.6 X 47 cm) equilibrated with buffer A. Protein was eluted with buffer A (0.5 ml/ min), and column fractions (3.1 ml) containing greater than 15 units of OMP decarboxylase activity/ml were pooled. This pool (16.5 ml, five fractions plus rinse) was loaded onto a DEAE-Bio-Gel A column (2.5 X 45 cm) equilibrated with buffer A. Protein was eluted (0.5 ml/ min) with a linear gradient of 0-0.3 M NaCl in 700 ml of buffer A. Column fractions (5.5 ml) containing greater than 15 units of OMP decarboxylase activity/ml were pooled. This pool (22.0 ml) was diluted by the addition of 2 volumes of buffer A and reapplied to the DEAE-Bio-Gel A column re-equilibrated with this buffer. Protein was eluted as described for the first DEAE-Bio-Gel A column. The five peak fractions, each of which contained greater than 5 units of OMP decarboxylase activity/ml, were pooled (26.5 ml). This purified sample was used for the analysis of OMP decarboxylase.
The purification procedure outlined above was developed after several unsuccessful attempts to purify OMP decarboxylase using the published protocol of Donovan and Kushner (1). In our hands, enzyme activity was not recovered from the blue Sepharose affinity column employed in this protocol. Negative results were obtained using blue Sepharose that had been prepared as described by Donovan and Kushner or purchased from Pharmacia P-L Biochemicals. One additional note: in contrast to previously reported results ( l ) , the specific activity of OMP decarboxylase in plasmid pDK26-containing cells grown on Luria broth was approximately one-tenth of that in the same strain grown on several minimal media (even when supplemented with 1% casamino acids and 1 mM uracil).
Determination of the Amino-terminal Amino Acid Sequence and Amino Acid Composition of OMP Decarboxylase-The second DEAE-Bio-Gel A column pool was dialyzed against 10 mM NH~HCOB and then dried in UQCUO. The sample was dissolved in 0.2 ml of water and redried in UQCUO seven times to remove residual NH,HCO,. A 0.1-mg sample (3.8 nmol of OMP decarboxylase subunit) was applied to an Applied Biosystems Model 470A gas-phase protein sequenator to determine the amino-terminal sequence. Phenylthiohydantoin-derivatives were identified by reverse-phase high-performance liquid chromotography. The amino acid composition was determined by hydrolyzing a 0.1-mg sample of purified protein in 2 ml of 6 N (constant boiling) HCl at 108 "C for 24 h. After evaporation in UUCUO, the sample was redissolved in water and subjected to automated amino acid analysis as described (21). A factor for converting nanomoles/milliliter amino acid to residues/OMP decarboxylase subunit was calculated from the recoveries of 10 relatively acid-stable amino acids (i.e. Asp, Thr, GIu, GIy, Ala, Leu, Phe, Lys, His, and Arg). The nanomoles of each of these amino acids c/ml was divided by the number of residues of that amino acid/subunit predicted from the DNA sequence. The conversion factor was the average of these quotients.

RESULTS
Identification and Nucleotide Sequence of the pyrF Structural Gene-The sequence of the pyrF structural gene and flanking regions was determined using DNA from plasmid pDK26. This pyrF-complementing plasmid contains a 2.2-kb insert of E. coli K12 chromosomal DNA. A 1549-bp region at one end of this insert is shown in Fig. 1. Previous subcloning experiments indicated that the PuuII site at nucleotide 582 in Fig. 1 is within thepyrF gene (3). Additional studies showed that the EcoRI site a t nucleotide 358 is between the pyrF promoter and all or nearly all of the pyrF structural gene (2,3). The pyrF promoter would be to the left of the EcoRI site as drawn in Fig. 1. The 1549-bp region was sequenced as shown in Fig. 1. The nucleotide sequence is shown in Fig. 2. Within the region predicted to contain the pyrF gene is an open reading frame that could encode a 27-kDa polypeptide, which is the size of the OMP decarboxylase subunit measured by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (1). This open reading frame begins with an ATG codon at nucleotide 320 and ends at nucleotide 1054 (Fig. 2). The ATG codon is preceded by a long Shine-Dalgarno sequence (SD,in Fig. 2) with spacing typical of a ribosome-binding site (22).
Translation of the open reading frame would result in the synthesis of a 245-amino acid polypeptide with a molecular weight of 26,350.
T o determine whether the identified open reading frame is the pyrF structural gene, OMP decarboxylase was purified and characterized. The purification procedure is described under "Materials and Methods" and summarized in Table I and Fig. 3. After the second DEAE-Bio-Gel A column, the enzyme was judged to be greater than 98% pure by scanning the protein bands in the gel shown in Fig. 3 with an LKB 2202 UltroScan laser densitometer. The subunit molecular mass was estimated to be 27 kDa (Fig. 3) as previously reported. The purified enzyme preparation was used to determine the amino-terminal amino acid sequence of OMP decarboxylase. The sequence of the first 32 amino acids was found to be identical to the sequence of residues 2-33 deduced from the DNA sequence of the open reading frame (Fig. 2). The purified enzyme was also used to determine the amino acid composition of OMP decarboxylase. As shown in Table 11, the determined composition and that predicted from the sequence of the open reading frame are in good agreement. These results establish that the open reading frame is the pyrF structural gene. Apparently the amino-terminal methionine residue of OMP decarboxylase is cleaved in vivo. Additional evidence supporting the location of the carboxylterminal end of the pyrF structural gene was provided by a codon preference plot of the 1549-bp sequence (data not shown). This computer-generated plot locates E. coli genes by virtue of their nonrandom codon usage (25). The plot showed that within the pyrF reading frame, nonrandom codon usage stops immediately after the termination codon of the structural gene. Codon usage in pyrF is typical of an E. coli protein coding sequence (26).
Mapping the pyrF Promoter and Transcriptional Regulntion-The pyrF promoter was located by primer extension mapping the 5' termini of pyrF transcripts as described under "Materials and Methods." Cellular RNA used as a source of pyrF transcripts was isolated from the pyrimidine-auxotrophic strain CLT43/pDK26 grown at 37 "C with either uracil or UMP as the sole pyrimidine source (doubling times were 48 and 54 min on uracil and UMP, respectively). Uracil is a good pyrimidine source and causes repressed pyr gene expression, whereas UMP is only slowly utilized and causes derepressed pyr gene expression. The DNA used as primer was the HincIIIEcoRI restriction fragment which contains bp 328-362 in Fig. 2. This fragment was 5'-32P-labeled at the EcoRI end. The DNA fragments extended by reverse transcriptase were analyzed on a sequencing gel containing a Maxam and Gilbert (15) sequencing ladder of the 360-bp RsaIIEcoRI restriction fragment shown in Fig. 1 that was 5'-32P-end-labeled at the same site as the primer (Fig. 4). The extended fragments were aligned with the fragments in the ladder that possess the same sequence after correcting for the slower migration by 1.5 nucleotides of the extended fragments. This correction is necessary because of the different 3' ends generated by reverse transcriptase and chemical cleavage in the sequencing reactions (27). The data indicate that in cells grown on uracil or UMP there is one major pyrF transcript that is initiated at the G residue at position 269 in Fig. 2. (Note that the sequences shown in Figs. 2 and 4 are of opposite strands.) A minor pyrF transcript apparently begins at the C residue at position 268. These transcription initiation sites are preceded by a -10 and a -35 sequence that are typical of an E. coli promoter (Fig. 2). The only other readily detectable extended fragment in Fig. 4 (lunes 1 and 2) would correspond to a minor transcript initiated at position 278. It is unlikely the transcription initiation occurs at this site because there is no upstream promoter-like sequence. It appears more likely that this third extended fragment is the result of premature termination by reverse transcriptase, perhaps caused by the region of dyad symmetry adjacent to nucleotide 278 (nucleotides 282-297 in Fig. 2).
The data shown in Fig. 4 indicate that the level of pyrF transcripts is higher in cells grown under derepressing conditions, which suggests that pyrF expression is regulated at the transcriptional level. To demonstrate transcriptional control of pyrF expression, the level of pyrF transcripts and of OMP decarboxylase were measured in strain CLT43/pDK26 grown at 30 "C with either uracil or UMP as the pyrimidine source. Cells were grown a t 30 "C instead of 37 "C to achieve a slightly higher level of derepression of pyrF expression (data not shown). The level of pyrF transcripts was measured by quantitative primer extension mapping as shown in Fig. 5. The results indicate that the level of pyrF transcripts in cells grown on UMP was 3-fold greater than that in cells grown on The nucleotide sequence shown is the antisense strand of the E. coli chromosomal fragment presented in Fig.  1; numbering is from the 5' end. The -10 and -35 hexamers of the pyrF promoter and the Shine-Dalgarno sequences for pyrF (SO,) and orfF (SDO) are underlined and labeled. The asterisk above nucleotide 269 indicates the primary transcription initiation site. Dyad symmetries are indicated by the arrows with the center of symmetry shown by the dots. Note that the shorter dyad symmetry contains a G-T base pair.  This step was included to reduce the level of trace contaminants.

CACAAAMGC CTGCCAGGGG ACAAATCGCA A E M T T TTTTATTTCC A C C G G m G C T C G C C G T TTACCTGTTT CGCGCCACTT
uracil (compare lanes 2 and 3 in Fig. 5). The level of OMP decarboxylase was measured in the same cultures used to quantitate pyrF transcripts and was also found t o be %fold higher in the cells grown on UMP (Table 111). These results clearly indicate transcriptional control. It should be noted, however, that the -fold derepression in strain CLT43/pDK26 was approximately half of that measured in strain CLT43, which carries only the chromosomal pyrF gene (Table 111).
The reason for the lower level of derepression in strain CLT43/pDK26 is unclear, but it was not due to a lower plasmid copy number in cells grown on UMP. The plasmid copy number in the cells grown on uracil or UMP was the same (&IO%). Nucleotide Sequence of the orfF Structural Gene-The nucleotide sequence downstream of pyrF was inspected for the orfF structural gene. This gene was previously shown to be cotranscribed with pyrF (2) and to map within the sequence shown in Fig. 2 (2,3). The orfF gene was also shown t o encode a polypeptide with a molecular mass of 13-15 kDa as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (2,3).  (Table I)   is not preceded by an appropriately spaced Shine-Dalgarno sequence that would function as a ribosome-binding site (22), but an adjacent methionine codon at bp 1057-1059 is preceded by such a sequence (SDo in Fig. 2). Examination of the entire open reading frame indicates that this second methionine codon, which overlaps the tandem pyrF termination codons, is the most likely site for translational initiation to occur. Translation starting at this site would result in a 108-amino Shown is an autoradiogram of a 6% polyacrylamide sequencing gel used to analyze DNA primers extended by reverse transcriptase. The primer extension reactions were as described under "Materials and Methods" with 0.5 pg (28,000 dpm) of primer DNA (see text) used in each reaction. The cellular RNA used was 31 pg from strain CLT43/ pDK26 grown on uracil (lane I ) or 27 pg from the same strain grown on UMP (lane 2). A control reaction in which no cellular RNA was included is shown in lune 3. The same amount of each primer extension reaction mixture was loaded on the gel. The Maxam and Gilbert (15) sequencing ladder was generated by sequencing the 360bp RsuI/EcoRI restriction fragment (bp 3-362 in Fig. 2) that had been 5'-"P-labeled at the EcoRI end, which is the same site labeled in the primer DNA. Nucleotide positions are numbered in accordance with the numbering of the complementary sequence shown in Fig. 2. acid polypeptide (Fig. 2) with a molecular mass of 11,396 daltons, which is only slightly less than that estimated for the orfF gene product. The proposed carboxyl-terminal end of the orfF structural gene coincides with the end of typical E. coli codon usage as indicated by a codon preference plot (data not shown). The end of the orfF structural gene is also closely followed by a region of dyad symmetry (bp 1389-1414 in Fig.  2), which may be involved in transcriptional termination (28). A computer search of the current GenBank nucleotide sequence (29) and the National Biomedical Research Foundation protein sequence (30) data libraries did not reveal sequences with extensive homology to the proposed orfF gene or its gene product.

Comparison of OMP Decarboxylase Amino Acid Sequences-
The amino acid sequence of the E. coli OMP decarboxylase pyrF Operon of E.

CLT43" Uracil
Cultures were grown a t 30 "C. Doubling times were 73 and 105 min on uracil and UMP, respectively, bThe numbers in parentheses indicate -fold derepression. The same -fold derepression was obtained by comparing total activities (data not shown).
was compared with the published sequences for this enzyme from S. cereuisiae (31), N . crmsa (32), and Ehrlich ascites cells (33) (Fig. 6). All amino acid sequences were deduced from nucleotide sequences. The Ehrlich ascites sequence represents the carboxyl-terminal domain of the bifunctional enzyme UMP synthase, which catalyzes both the fifth and sixth steps in the de nouo synthesis of UMP (33). The conlparison revealed two major features. First, throughout the E. coli sequence there are regions of obvious homology with the eucaryotic sequences. There are 40 positions with identical amino acids in each sequence and a number of other sites where all 4 residues are functionally similar. Second, the eucaryotic sequences are much more closely related to each other than to the E. coli sequence. There are 55 positions a t which only the eucaryotic residues are identical, but only 21 positions where the E. coli and 2 eucaryotic residues are the same. The most highly conserved region in the four enzymes appears to be that corresponding to E. coli residues 66-80, which is therefore likely to include a site that is essential for enzymatic activity.

DISCUSSION
In E. coli and Salmonella typhimurium, the expression of the genes and operons that encode the six pyrimidine biosynthetic enzymes is noncoordinately regulated by the intracellular levels of uridine and cytidine nucleotides (4, 34). The regulation of pyrF expression resembles that of pyrBI and pyrE in that a uridine nucleotide appears to be the sole pyrimidine effector. In the case ofpyrBI andpyrE, the effector is known to be U T P (3.536). Recent studies have shown that pyrBI andpyrE expression is regulated by similar attenuation control mechanisms (6-13). In these mechanisms, transcriptional termination a t a p-independent terminator (attenuator) immediately preceding the pyr structural gene(s) is regulated by the relative rates of UTP-sensitive transcription and coupled translation within a leader region upstream of the attenuator.
A primary objective of the present study was to examine the region between the pyrF promoter and structural gene for sequences similar to the regulatory elements found in the pyrBI and pyrE leader regions, i.e. a UTP-sensitive transcriptional pause site and a p-independent terminator. No such sequences were found (not even in the structural gene), which clearly indicates that the mechanism regulating pyrF expression is different than that for pyrBI and pyrE. Inspection of the pyrF leader sequence did reveal one region of hyphenated dyad symmetry (nucleotides 282-297 in Fig. 2); however, its regulatory significance is unknown.
The quantitative primer extension mapping data presented in Fig. 5 also exclude the possibility that pyrF expression is regulated by a p-dependent attenuation control mechanism.
The results indicate that regulation occurs a t a step before the transcription of the 11th codon of the pyrF structural gene. (Note that the oligonucleotide primer used to quantitate pyrF transcripts in this experiment is complementary to codons 2-10.) p-Dependent termination could not occur within this early stage of pyrF transcription because the 5' end of the transcript is too short for p binding (28).
The results presented indicate that pyrF expression is regulated at the transcriptional level. The levels of pyrF transcripts and OMP decarboxylase are derepressed coordinately in cells grown under conditions of pyrimidine limitation. The -fold derepression of pyrF expression measured with strain CLT43/pDK26, which contains multiple copies of the pyrF operon, was approximately half of that observed with untransformed strain CLT43. This result may indicate the involvement of a titratable regulatory element like a DNA-binding protein. The identification of such a protein, if one actually exists, as well as other regulatory elements will require additional experiments.
The orfF structural gene was tentatively identified within the nucleotide sequence immediately downstream of the pyrF structural gene. In fact, the putative orfF translational initiation codon overlaps the pyrF termination codon. This overlap may indicate translational coupling between these two cotranscribed genes. The predicted molecular mass of the orfF gene product, which is 11.4 kDa, is approximately 2 kDa less FIG. 6. Comparison of the OMP decarboxylase amino acid sequences of E. coli, S. cerevisiae. N. crassa, and Ehrlich ascites cells (mouse). The sequences were aligned with the aid of the BESTFIT computer program supplied by the University of Wisconsin Genetics Computer Group to yield maximum homology with respect to identical and functionally similar amino acids. Identical residues at a given position are enclosed in boxes. Positions at which all residues are functionally similar are indicated by asterisks. Two long nonhomologous regions in the N . crassa sequence which include residues 152-243 and 297-314 are not shown in the figure.
Numbering begins at the amino-terminal residue of each sequence.  . . . . than that estimated from the relative mobility of the polypeptide during sodium dodecyl sulfate-polyacrylamide gel electrophoresis. This difference is probably within the experimental error for determining the size of a small protein using this procedure (37). If the orfF gene product were, in fact, the same size as that measured, the orfF structural gene would have to begin within the carboxyl-terminal portion of the pyrF structural gene. There are no obvious translational initiation sites in this region of structural gene overlap, but an atypical translational initiation event cannot be excluded at present. The apparent coupling of pyrF and orfF expression is of particular interest because it suggests that the orfF gene product may be involved in pyrimidine nucleotide metabolism. Future experiments will be directed a t identifying the physiological role of this protein.

A V T V L T S H E A S D L V D L G M T L S P A D Y A G I V S G L K Q A A E E V T K E
The availability of the amino acid sequence of the E. coli OMP decarboxylase provided an opportunity to compare procaryotic and eucaryotic sequences of this enzyme. The E. coli sequence is the only procaryotic OMP decarboxylase sequence presently known. It was compared to the sequences of the enzymes from S. cereuisiue, N . crassa, and Ehrlich ascites cells (see ''Results"). This comparison should facilitate the identification of amino acids that play an important role in both the shared and unique properties of these enzymes.