Phase I clinical trial of the Src inhibitor dasatinib with dacarbazine in metastatic melanoma.

BACKGROUND: Src inhibitors sensitise melanoma cells to chemotherapy in preclinical models. The combination of dasatinib and dacarbazine was tested in a phase I trial in melanoma. METHODS: Patients had ECOG performance status 0-2 and normal organ function. Dacarbazine was administered on day 1 and dasatinib on day 2 through 19 of each 21-day cycle. Both were escalated from 50 mg b.i.d. of dasatinib and 800 mg m(-2) of dacarbazine. Available pre-treatment biopsies were sequenced for BRAF, NRAS, and C-Kit mutations. RESULTS: Dose-limiting toxicity was reached at dasatinib 70 mg b.i.d./dacarbazine 1000 mg m(-2), and was predominantly haematological. In 29 patients receiving dasatinib 70 mg b.i.d., the objective response rate (ORR) was 13.8%, the clinical benefit rate (ORR+SD) was 72.4%, the 6-month progression-free survival (PFS) was 20.7%, and the 12-month overall survival (OS) was 34.5%. Two out of three patients who were wild type for BRAF, NRAS, and c-KIT mutations had confirmed partial responses, and one had a minor response. CONCLUSION: The recommended phase II dose is dasatinib 70 mg b.i.d with dacarbazine 800 mg m(-2). PFS and OS data for dasatinib at 70 mg b.i.d. with dacarbazine compared favourably with historical controls. Preliminary data support evaluating tumour mutation status further as a biomarker of response.

Caregiver burden and symptom distress in people with cancer receiving hospice care.

PURPOSE/OBJECTIVES: To examine the relationship between caregiver burden and symptom distress in patients with terminal cancer who are enrolled in hospice. DESIGN: Descriptive, quantitative. SETTING: A large, metropolitan, nonprofit-based organization in west central Florida. SAMPLE: Convenience sample of 30 patient-caregiver dyads enrolled in hospice. METHODS: Caregivers completed the Caregiver Reaction Scale to measure the level of caregiver burden; patients completed the Adopted Symptom Distress Scale. Results were correlated using a Pearson correlation. MAIN RESEARCH VARIABLES: Symptom distress and caregiver burden. FINDINGS: The patient sample exhibited low symptom distress, and the caregiver sample exhibited moderate caregiver burden. A statistically significant moderate correlation existed between symptom distress and caregiver burden. CONCLUSIONS: The significant moderate correlation confirms the idea that caregiver burden and patient symptom distress are related. Future studies are needed to obtain a more representative sample of caregivers of patients closer to death, even if those patients are nonresponsive. IMPLICATIONS FOR NURSING PRACTICE: This information can assist hospice nurses in assessing and formulating targeted care for symptom distress and caregiver burden in their patients,

Iron acquisition and virulence in Helicobacter pylori: a major role for FeoB, a high-affinity ferrous iron transporter.

The genome sequence of Helicobacter pylori suggests that this bacterium possesses several Fe acquisition systems, including both Fe2+- and Fe3+-citrate transporters. The role of these transporters was investigated by generating insertion mutants in feoB, tonB, fecA1 and fecDE. Fe transport in the feoB mutant was approximately 10-fold lower than in the wild type (with 0.5 microM Fe), irrespective of whether Fe was supplied in the Fe2+ or Fe3+ form. In contrast, transport rates were unaffected by the other mutations. Complementation of the feoB mutation fully restored both Fe2+ and Fe3+ transport. The growth inhibition exhibited by the feoB mutant in Fe-deficient media was relieved by human holo-transferrin, holo-lactoferrin and Fe3+-dicitrate, but not by FeSO4. The feoB mutant had less cellular Fe and was more sensitive to growth inhibition by transition metals in comparison with the wild type. Biphasic kinetics of Fe2+ transport in the wild type suggested the presence of high- and low-affinity uptake systems. The high-affinity system (apparent Ks = 0.54 microM) is absent in the feoB mutant. Transport via FeoB is highly specific for Fe2+ and was inhibited by FCCP, DCCD and vanadate, indicating an active process energized by ATP. Ferrozine inhibition of Fe2+ and Fe3+ uptake implied the concerted involvement of both an Fe3+ reductase and FeoB in the uptake of Fe supplied as Fe3+. Taken together, the results are consistent with FeoB-mediated Fe2+ uptake being a major pathway for H. pylori Fe acquisition. feoB mutants were unable to colonize the gastric mucosa of mice, indicating that FeoB makes an important contribution to Fe acquisition by H. pylori in the low-pH, low-O2 environment of the stomach.

Ferritin mutants of Escherichia coli are iron deficient and growth impaired, and fur mutants are iron deficient.

Escherichia coli contains at least two iron storage proteins, a ferritin (FtnA) and a bacterioferritin (Bfr). To investigate their specific functions, the corresponding genes (ftnA and bfr) were inactivated by replacing the chromosomal ftnA and bfr genes with disrupted derivatives containing antibiotic resistance cassettes in place of internal segments of the corresponding coding regions. Single mutants (ftnA::spc and bfr::kan) and a double mutant (ftnA::spc bfr::kan) were generated and confirmed by Western and Southern blot analyses. The iron contents of the parental strain (W3110) and the bfr mutant increased by 1.5- to 2-fold during the transition from logarithmic to stationary phase in iron-rich media, whereas the iron contents of the ftnA and ftnA bfr mutants remained unchanged. The ftnA and ftnA bfr mutants were growth impaired in iron-deficient media, but this was apparent only after the mutant and parental strains had been precultured in iron-rich media. Surprisingly, ferric iron uptake regulation (fur) mutants also had very low iron contents (2.5-fold less iron than Fur+ strains) despite constitutive expression of the iron acquisition systems. The iron deficiencies of the ftnA and fur mutants were confirmed by Mössbauer spectroscopy, which further showed that the low iron contents of ftnA mutants are due to a lack of magnetically ordered ferric iron clusters likely to correspond to FtnA iron cores. In combination with the fur mutation, ftnA and bfr mutations produced an enhanced sensitivity to hydroperoxides, presumably due to an increase in production of "reactive ferrous iron." It is concluded that FtnA acts as an iron store accommodating up to 50% of the cellular iron during postexponential growth in iron-rich media and providing a source of iron that partially compensates for iron deficiency during iron-restricted growth. In addition to repressing the iron acquisition systems, Fur appears to regulate the demand for iron, probably by controlling the expression of iron-containing proteins. The role of Bfr remains unclear.

Iron storage in bacteria.

Iron is an essential nutrient for nearly all organisms but presents problems of toxicity, poor solubility and low availability. These problems are alleviated through the use of iron-storage proteins. Bacteria possess two types of iron-storage protein, the haem-containing bacterioferritins and the haem-free ferritins. These proteins are widespread in bacteria, with at least 39 examples known so far in eubacteria and archaebacteria. The bacterioferritins and ferritins are distantly related but retain similar structural and functional properties. Both are composed of 24 identical or similar subunits (approximately 19 kDa) that form a roughly spherical protein (approximately 450 kDa, approximately 120 A diameter) containing a large hollow centre (approximately 80 A diameter). The hollow centre acts as an iron-storage cavity with the capacity to accommodate at least 2000 iron atoms in the form of a ferric-hydroxyphosphate core. Each subunit contains a four-helix bundle which carries the active site or ferroxidase centre of the protein. The ferroxidase centres endow ferrous-iron-oxidizing activity and are able to form a di-iron species that is an intermediate in the iron uptake, oxidation and core formation process. Bacterioferritins contain up to 12 protoporphyrin IX haem groups located at the two-fold interfaces between pairs of two-fold related subunits. The role of the haem is unknown, although it may be involved in mediating iron-core reduction and iron release. Some bacterioferritins are composed of two subunit types, one conferring haem-binding ability (alpha) and the other (beta) bestowing ferroxidase activity. Bacterioferritin genes are often adjacent to genes encoding a small [2Fe-2S]-ferredoxin (bacterioferritin-associated ferredoxin or Bfd). Bfd may directly interact with bacterioferritin and could be involved in releasing iron from (or delivering iron to) bacterioferritin or other iron complexes. Some bacteria contain two bacterioferritin subunits, or two ferritin subunits, that in most cases co-assemble. Others possess both a bacterioferritin and a ferritin, while some appear to lack any type of iron-storage protein. The reason for these differences is not understood. Studies on ferritin mutants have shown that ferritin enhances growth during iron starvation and is also involved in iron accumulation in the stationary phase of growth. The ferritin of Campylobacter jejuni is involved in redox stress resistance, although this does not appear to be the case for Escherichia coli ferritin (FtnA). No phenotype has been determined for E. coli bacterioferritin mutants and the precise role of bacterioferritin in E. coli remains uncertain.

Interaction of nitric oxide with non-haem iron sites of Escherichia coli bacterioferritin: reduction of nitric oxide to nitrous oxide and oxidation of iron(II) to iron(III).

The bacterioferritin (BFR) of Escherichia coli consists of 24 identical subunits, each containing a dinuclear metal-binding site consisting of two histidines and four carboxylic acid residues. Earlier studies showed that the characterization of iron binding to BFR could be aided by EPR analysis of iron-nitrosyl species resulting from the addition of NO to the protein [Le Brun, Cheesman, Andrews, Harrison, Guest, Moore and Thomson (1993) FEBS Lett. 323, 261-266]. We now report data from gas chromatographic head space analysis combined with EPR spectroscopy to show that NO is not an inert probe: iron(II)-BFR catalyses the reduction of NO to N2O, resulting in oxidation of iron(II) at the dinuclear centre and the subsequent detection of mononuclear iron(III). In the presence of excess reductant (sodium ascorbate), iron(II)-BFR also catalyses the reduction of NO to N2O, giving rise to three mononuclear iron-nitrosyl species which are detectable by EPR. One of these, a dinitrosyl-iron complex of S = 1/2, present at a maximum of one per subunit, is shown by EPR studies of site-directed variants of BFR not to be located at the dinuclear centre. This is consistent with a proposal that the diferric form of the centre is unstable and breaks down to form mononuclear iron species.

Spectroscopic and voltammetric characterisation of the bacterioferritin-associated ferredoxin of Escherichia coli.

The bacterioferritin-associated ferredoxin (Bfd) of Escherichia coli is a 64-residue polypeptide encoded by the bfd gene located upstream of the gene (bfr) encoding the iron-storage haemoprotein, bacterioferritin. The Bfd sequence resembles those of the approximately 60-residue domains found in NifU proteins (required for metallocluster assembly), nitrite reductases, and Klebsiella pneumoniae nitrate reductase. These related-domains contain four well-conserved cysteine residues, which are thought to function as ligands to a [2Fe-2S] cluster. The Bfd protein was over-produced, purified, and characterised. Bfd was found to be a positively-charged monomer containing two iron atoms and two labile sulphides. Ultraviolet-visible, EPR, variable-temperature magnetic-circular dichroism and resonance Raman spectroscopies, together with cyclic voltogram measurements, revealed the presence of a [2Fe-2S]2+,+ centre (E1/2 = -254 mV) having remarkably similar properties to the Fe-S cluster of NifU. Bfd may thus be a 2Fe ferredoxin participating either in release/delivery of iron from/to bacterioferritin (or other iron complexes), or in iron-dependent regulation of bfr expression.

Iron incorporation into ferritins: evidence for the transfer of monomeric Fe(III) between ferritin molecules and for the formation of an unusual mineral in the ferritin of Escherichia coli.

Iron that has been oxidized by H-chain ferritin can be transferred into other ferritin molecules before it is incorporated into mature ferrihydrite iron cores. Iron(III) dimers are formed at the ferroxidase centres of ferritin H chains at an early stage of Fe(II) oxidation. Mössbauer spectroscopic data now show that the iron is transferred as monomeric species arising from dimer dissociation and that it binds to the iron core of the acceptor ferritin. Human H-chain ferritin variants containing altered threefold channels can act as acceptors, as can the ferritin of Escherichia coli (Ec-FTN). A human H-chain ferritin variant with a substituted tyrosine (rHuHF-Y34F) can act as a donor of Fe(III). Since an Fe(III)-tyrosinate (first identified in bullfrog H-chain ferritin) is absent from variant rHuHF-Y34F, the Fe(III) transferred is not derived from this tyrosinate complex. Mössbauer parameters of the small iron cores formed within Ec-FTN are significantly different from those of mammalian ferritins. Analysis of the spectra suggests that they are derived from both ferrihydrite and non-ferrihydrite components. This provides further evidence that the ferritin protein shell can influence the structure of its iron core.

Direct observation of the iron binding sites in a ferritin.

X-Ray analysis of the ferritin of Escherichia coli (Ec-FTN) and of Ec-FTN crystals soaked in (NH4)2Fe(SO4)2 has revealed the presence of three iron-binding sites per subunit. Two of these form a di-iron site in the centre of the subunit as has been proposed for the 'ferroxidase centres' of human ferritin H chains. This di-iron site, lying within the 4-alpha-helix bundle, resemble those of ribonucleotide reductase, methane monoxygenase and haemerythrin. The third iron is bound by ligands unique to Ec-FTN on the inner surface of the protein shell. It is speculated that this state may represent the nucleation centre of a novel type of Fe(III) cluster, recently observed in Ec-FTN.

An EPR investigation of non-haem iron sites in Escherichia coli bacterioferritin and their interaction with phosphate. A study using nitric oxide as a spin probe.

EPR studies of bacterioferritin (BFR), an iron-storage protein of Escherichia coli [1993, Biochem. J. 292, 47-56], have revealed the presence of non-haem iron (III) (NHI) sites within the protein coat which may be involved in iron uptake and release. When nitric oxide was used as an EPR spin probe of the Fe(II) state of the NHI sites, two distinct mononuclear NHI species were found. Under certain conditions, an iron dimer was also observed. The reaction of phosphate with NHI species has been investigated. Results point to a function for this anion in core nucleation.

Haem and non-haem iron sites in Escherichia coli bacterioferritin: spectroscopic and model building studies.

The bacterioferritin (BFR) of Escherichia coli is an iron-storage protein containing 24 identical subunits and between three and 11 protohaem IX groups per molecule. Titration with additional haem gave a maximum loading of 12-14 haems per molecule. The e.p.r. spectra and magnetic c.d. spectra of the protein-bound haem show it to be low-spin Fe(III), and coordinated by two methionine residues as previously reported for BFRs isolated from Pseudomonas aeruginosa and Azotobacter vinelandii [Cheesman, Thomson, Greenwood, Moore and Kadir, Nature (London) (1990) 346, 771-773]. A recent sequence alignment indicated that BFR may be structurally related to ferritin. The molecular model proposed for E. coli BFR has a four-alpha-helix-bundle subunit conformation and a quaternary structure similar to those of mammalian ferritins. In this model there are two types of hydrophobic pocket within which two methionine residues are correctly disposed to bind haem. The e.p.r. spectra also reveal a monomeric non-haem Fe(III) species with spin, S = 5/2. On the basis of sequence comparisons, a ferroxidase centre has recently been proposed to be present in BFR [Andrews, Smith, Yewdall, Guest and Harrison (1991) FEBS Lett. 293, 164-168] and the possibility that this Fe(III) ion may reside at or near the ferroxidase centre is discussed.

Overproduction, purification and characterization of the bacterioferritin of Escherichia coli and a C-terminally extended variant.

The bacterioferritin (BFR) of Escherichia coli is an iron-sequestering haemoprotein composed of 24 identical polypeptide chains forming an approximately spherical protein shell with a central iron-storage cavity. BFR and BFR-lambda, a variant with a 14-residue C-terminal extension, have been amplified (120-fold and 50-fold, respectively), purified by a new procedure and characterized. The overproduced BFR exhibited properties similar to those of natural BFR, but the iron content (25-75 non-haem Fe atoms/molecule) was 13-39-fold lower. Two major assembly states of BFR were detected, a 24-subunit protein (tetracosamer) and a novel haem-containing subunit dimer. BFR-lambda subunits assembled into tetracosamers having the same external-surface properties as BFR, presumably because their C-terminal extensions project into and occupy about 60% of the central cavity. As a result, BFR-lambda failed totake up iron under conditions that allowed incorporation into BFR in vitro. The haem content of BFR-lambda (1-2 haems/tetracosamer) was lower than that of BFR (3.5-10.5 haems/tetracosamer) and this, together with a difference in the visible spectra of the two haemoproteins, suggested that the C-terminal extensions in BFR-lambda perturb the haem-binding pockets. A subunit dimer form of BFR-lambda was not detected. A combination of Mössbauer spectroscopy and electron diffraction showed that the BFR loaded with iron in vitro has a ferrihydrite-like iron core, whereas the in-vivo loaded protein has an amorphous core.

Treatment of cancers involving the liver and porta hepatis with external beam irradiation and intraarterial hepatic fluorodeoxyuridine.

A Phase I/II clinical trial was designed for patients with malignancies of the liver and porta hepatis. This protocol employed three concepts: a) boost treatment to gross tumor within the liver for selected patients, determined by the dose-volume histogram (DVH) of the normal liver that would be irradiated by boost treatment; b) concurrent use of intraarterial hepatic 5-fluorodeoxyuridine (FdUrd) as a radiosensitizer; and c) hyperfractionation (1.5 Gy fractions given bid greater than 4 hr apart). This report describes the results of treatment of the first 33 patients entered onto this study, with a minimum follow-up of 1 year. Twenty patients received only whole liver irradiation (33 Gy). Thirteen patients were treated with whole liver irradiation (30 Gy) plus a 15 Gy (6 patients) or 30 Gy (7 patients) boost (total 45 Gy and 60 Gy to the tumor, respectively). Forty-eight percent of the evaluable patients (14/29) had an objective response, based on CT scan. The median duration of response was 8 months. The chief toxicities were fatigue, nausea, gastritis, and diarrhea, which were less than or equal to grade 2 in severity. Two patients developed mild radiation hepatitis which was treated successfully with diuretics. These data suggest that the treatment of intrahepatic malignancies can be guided by the concept of DVH analysis of the normal liver to allow the safe administration of doses of radiation that are potentially tumoricidal and are well above those that would be predicted to be tolerable for the whole liver.

Long-term central venous access with a peripherally placed subcutaneous infusion port: initial results.

A new subcutaneous infusion port and catheter system for long-term central venous access, designed to be implanted in the interventional radiology suite, was evaluated. In 35 patients, a 5-F polyurethane catheter was placed in the superior vena cava via the axillary or brachial venous approach under fluoroscopic guidance. A 2.5 X 2.5-cm2 subcutaneous pocket was dissected for the port. The port was then connected to the catheter, and the incision was closed. Ports have been implanted for a total of 5,290 patient days (5-307 days for an individual patient). Blood transfusion, bolus drug administrations, and 5-day outpatient chemotherapy infusions were successful in all attempts. Blood sampling was successful in 98.9% of attempts. No infectious or thrombotic complications were encountered. Acceptance of this device by patients and nursing staff has been excellent. The initial results indicate that this peripherally placed port is a viable alternative for patients requiring long-term central venous access.

Modulation of liver tumor blood flow with hepatic arterial epinephrine: a SPECT study.

Changes in the relative arterial flow to hepatic tumors and adjacent normal liver, in response to varied doses of hepatic arterial epinephrine, were studied with single photon emission computed tomography. In 18 patients with known hepatic tumors, hepatic artery perfusion scans were obtained with the concurrent infusion of technetium-99m-labeled macroaggregated albumin and escalating doses of epinephrine (0-10 micrograms/min). Regions of interest were drawn around tumor and adjacent normal liver in three planes, and the average tumor-to-liver ratio (T:L) was calculated. In all 18 patients, there was a measurable baseline T:L perfusion advantage (range, 1.7-18.7; mean, 4.8). In 12 of 18 patients, this ratio increased with epinephrine (range, 1.1-53.6 times the baseline value; mean, 7.1). In six patients, no improvement in T:L could be demonstrated. In 14 patients the lung shunt index, a measurement of arteriovenous shunting, increased with escalating doses of epinephrine. This pilot study suggests that the infusion of epinephrine may improve the therapeutic index of certain regional therapies such as bolus drug infusions, hepatic arterial embolization, and radioactive microsphere therapy.

Cloning, sequencing, and mapping of the bacterioferritin gene (bfr) of Escherichia coli K-12.

The bacterioferritin (BFR) of Escherichia coli K-12 is an iron-storage hemoprotein, previously identified as cytochrome b1. The bacterioferritin gene (bfr) has been cloned, sequenced, and located in the E. coli linkage map. Initially a gene fusion encoding a BFR-lambda hybrid protein (Mr 21,000) was detected by immunoscreening a lambda gene bank containing Sau3A restriction fragments of E. coli DNA. The bfr gene was mapped to 73 min (the str-spc region) in the physical map of the E. coli chromosome by probing Southern blots of restriction digests of E. coli DNA with a fragment of the bfr gene. The intact bfr gene was then subcloned from the corresponding lambda phage from the gene library of Kohara et al. (Y. Kohara, K. Akiyama, and K. Isono, Cell 50:495-508, 1987). The bfr gene comprises 474 base pairs and 158 amino acid codons (including the start codon), and it encodes a polypeptide having essentially the same size (Mr 18,495) and N-terminal sequence as the purified protein. A potential promoter sequence was detected in the 5' noncoding region, but it was not associated with an "iron box" sequence (i.e., a binding site for the iron-dependent Fur repressor protein). BFR was amplified to 14% of the total protein in a bfr plasmid-containing strain. An additional unidentified gene (gen-64), encoding a relatively basic 64-residue polypeptide and having the same polarity as bfr, was detected upstream of the bfr gene.

Nucleotide sequence of the FNR-regulated fumarase gene (fumB) of Escherichia coli K-12.

The nucleotide sequence of a 3,162-base-pair (bp) segment of DNA containing the FNR-regulated fumB gene, which encodes the anaerobic class I fumarase (FUMB) of Escherichia coli, was determined. The structural gene was found to comprise 1,641 bp, 547 codons (excluding the initiation and termination codons), and the gene product had a predicted Mr of 59,956. The amino acid sequence of FUMB contained the same number of residues as did that of the aerobic class I fumarase (FUMA), and there were identical amino acids at all but 56 positions (89.8% identity). There was no significant similarity between the class I fumarases and the class II enzyme (FUMC) except in one region containing the following consensus: Gly-Ser-Xxx-Ile-Met-Xxx-Xxx-Lys-Xxx-Asn. Some of the 56 amino acid substitutions must be responsible for the functional preferences of the enzymes for malate dehydration (FUMB) and fumarate hydration (FUMA). Significant similarities between the cysteine-containing sequence of the class I fumarases (FUMA and FUMB) and the mammalian aconitases were detected, and this finding further supports the view that these enzymes are all members of a family of iron-containing hydrolyases. The nucleotide sequence of a 1,142-bp distal sequence of an unidentified gene (genF) located upstream of fumB was also defined and found to encode a product that is homologous to the product of another unidentified gene (genA), located downstream of the neighboring aspartase gene (aspA).

Combination chemo-radiation therapy for jaundice due to focal malignant obstruction of the major bile ducts.

Twenty patients with focal malignant obstruction of the major bile ducts (6 cholangiocarcinoma, 8 colorectal, 3 hepatoma, 2 unknown primary, and 1 gastric cancer) were treated on a protocol examining the toxicity and efficacy in relieving jaundice of external beam radiation therapy (4500 cGy in 300 cGy fractions) combined with continuous hepatic arterial (15 patients) or peripheral venous (5 patients) fluorouracil infusion. Toxicity of this regimen consisted of anorexia with mild nausea and vomiting in 55% of patients and gastric ulceration (responsive to medical management) in 15% of patients. One patient exhibited transient grade 2 hepatic toxicity and one had asymptomatic grade 4 leukopenia. Of 14 patients treated without prior biliary drainage, 8 exhibited a decrease in bilirubin levels from a mean of 14.5 mg/dl to 1.5 mg/dl. Four of six patients with biliary drainage catheters at the start of treatment were able to have them removed without reobstruction. For the 8 responding patients among those who did not have cholangiocarcinomas, the median response duration was 5 months with a median survival from treatment of 6.5 months. For the 4 responding patients with cholangiocarcinoma, the median response duration was 16 months with a median survival from treatment of 20 months. All responders did not have a return of jaundice due to reobstruction of the major ducts (until death or to the present). All responders who have died did so due to tumor progression outside of the treated field except for one who died of unrelated causes. The mean number of proven or presumed episodes of cholangitis per patient was virtually identical in those without (1.8) and those with stents/tubes (1.4, p = 0.561). This regionally focused combined modality cytotoxic therapy was able to relieve obstruction in the majority of patients without excess morbidity (including a lack of any detectable increase in sepsis). Thus, it appears feasible to consider randomized studies of this cytotoxic approach versus standard mechanical drainage procedures to define the relative risks and benefits of each.

Nucleotide sequence of the gene encoding the GMP reductase of Escherichia coli K12.

(1) The nucleotide sequence of a 1991 bp segment of DNA that expresses the GMP reductase (guaC) gene of Escherichia coli K12 was determined. (2) This gene comprises 1038 bp, 346 codons (including the initiation codon but excluding the termination codon), and it encodes a polypeptide of Mr 37,437 which is in good agreement with previous maxicell studies. (3) The sequence contains a putative promoter 102 bp upstream of the translational start codon, and this is immediately followed by a (G + C)-rich discriminator sequence suggesting that guaC expression may be under stringent control (4) The GMP reductase exhibits a high degree of sequence identity (34%) with IMP dehydrogenase (the guaB gene product) indicative of a close evolutionary relationship between the salvage pathway and the biosynthetic enzymes, GMP reductase and IMP dehydrogenase, respectively. (5) A single conserved cysteine residue, possibly involved in IMP binding to IMP dehydrogenase, was located within a region that possesses some of the features of a nucleotide binding site. (6) The IMP dehydrogenase polypeptide contains an internal segment of 123 amino acid residues that has no counterpart in GMP reductase and may represent an independent folding domain flanked by (alanine + glycine)-rich interdomain linkers.