Antidepressant treatment of neuropathic pain: looking for the mechanism

Neuropathic pain arises as a direct consequence of a lesion or disease affecting the somatosensory system. Among the recommended first-line treatments are antidepressant drugs – that is, molecules that were initially developed to treat other disorders of the nervous system. While their clinical efficacy against neuropathic pain was established more than 30 years ago, there is little information on the mechanism underlying their antidepressant action. However, understanding the therapeutic mechanism of these treatments could help to improve them, or even lead to new therapeutic approaches. In this article, we discuss the difficulties in conducting relevant preclinical research on neuropathic pain treatment with antidepressant drugs and we present the most recent findings on the putative mechanism, which highlight the role of β2-adrenoceptors and δ-opioid receptors.

Chronic treatment with agonists of beta(2)-adrenergic receptors in neuropathic pain

Expression of beta(2)-adrenoceptors (beta(2)-ARs) within the nociceptive system suggested their potential implication in nociception and pain. Recently, we demonstrated that these receptors are essential for neuropathic pain treatment by antidepressant drugs. The aim of the present study was to investigate whether the stimulation of beta(2)-ARs could in fact be adequate to alleviate neuropathic allodynia. Neuropathy was induced in mice by sciatic nerve cuffing. We demonstrate that chronic but not acute stimulation of beta(2)-ARs with agonists such as clenbuterol, formoterol, metaproterenol and procaterol suppressed neuropathic allodynia. By using a pharmacological approach with the beta(2)-AR antagonist ICI 118,551 or a transgenic approach with mice deficient for beta(2)-ARs, we confirmed that the antiallodynic effect of these agonists was specifically related to their action on beta(2)-ARs. We also showed that chronic treatment with the beta(1)-AR agonist xamoterol or with the beta(3)-AR agonist BRL 37344 had no effect on neuropathic allodynia. Chronic stimulation of beta(2)-ARs, but not beta(1)- or beta(3)-ARs, by specific agonists is thus able to alleviate neuropathic allodynia. This action of beta(2)-AR agonists might implicate the endogenous opioid system; indeed chronic clenbuterol effect can be acutely blocked by the delta-opioid receptor antagonist naltrindole. Present results show that beta(2)-ARs are not only essential for the antiallodynic action of antidepressant drugs on sustained neuropathic pain, but also that the stimulation of these receptors is sufficient to relieve neuropathic allodynia in a murine model. Our data suggest that beta(2)-AR agonists may potentially offer an alternative therapy to antidepressant drugs for the chronic treatment of neuropathic pain.

Delta-opioid receptors are critical for tricyclic antidepressant treatment of neuropathic allodynia

BACKGROUND

The therapeutic effect of antidepressant drugs against depression usually necessitates a chronic treatment. A large body of clinical evidence indicates that antidepressant drugs can also be highly effective against chronic neuropathic pain. However, the mechanism by which these drugs alleviate pain is still unclear.

METHODS

We used a murine model of neuropathic pain induced by sciatic nerve constriction to study the antiallodynic properties of a chronic treatment with the tricyclic antidepressants nortriptyline and amitriptyline. Using knockout and pharmacological approaches in mice, we determined the influence of delta-opioid receptors in the therapeutic action of chronic antidepressant treatment.

RESULTS

In our model, a chronic treatment by tricyclic antidepressant drugs totally suppresses the mechanical allodynia in neuropathic C57Bl/6J mice. This therapeutic effect can be acutely reversed by an injection of the delta-opioid receptor antagonist naltrindole. Moreover, the antiallodynic property of antidepressant treatment is absent in mice deficient for the delta-opioid receptor gene.

CONCLUSIONS

The antiallodynic effect of chronic antidepressant treatment is mediated by a recruitment of the endogenous opioid system acting through delta-opioid receptors.

Glia-induced neuronal differentiation by transcriptional regulation

There is increasing evidence that different phases of brain development depend on neuron-glia interactions including postnatal key events like synaptogenesis. To address how glial cells influence synapse development, we analyzed whether and how glia-derived factors affect gene expression in primary cultures of immunoisolated rat retinal ganglion cells (RGCs) by oligonucleotide microarrays. Our results show that the transcript pattern matched the developmental stage and characteristic properties of RGCs in vitro. Glia-conditioned medium (GCM) and cholesterol up- and downregulated a limited number of genes that influence the development of dendrites and synapses and regulate cholesterol and fatty acid metabolism. The oligonucleotide microarrays detected the transcriptional regulation of neuronal cholesterol homeostasis in response to GCM and cholesterol treatment. Surprisingly, our study revealed neuronal expression and glial regulation of matrix gla protein (Mgp). Together, our results suggest that glial cells promote different aspects of neuronal differentiation by regulating transcription of distinct classes of genes.

Diltiazem-induced neuroprotection in glutamate excitotoxicity and ischemic insult of retinal neurons

PURPOSE

Cell death is often related to an abnormal increase in Ca(2+) flux. In the retina, Ca(2+) channels are mainly from the L-type that do not inactivate with time. Under excitotoxic and ischemic conditions, their continuous activation may therefore contribute significantly to the lethal Ca(2+) influx. To assess this hypothesis, the Ca(2+) channel blocker, diltiazem, was applied in excitotoxic and ischemic conditions.

METHODS

To induce excitotoxicity, retinal cell cultures from newborn rats were incubated with glutamate. The toxicity of glutamate was quantified by neuronal immunostaining with an antibody directed against the neuron specific enolase. Glutamate receptor function in vitro was assessed in pig retinal cell cultures by patch clamp recording. Retinal ischemia was induced by raising the intraocular pressure in adult rats. Retinal cell loss was quantified on retinal sections by measuring nuclear cell densities.

RESULTS

In retinal cell culture, glutamate application induced a major cell loss. This cell loss was attributed to glutamate excitotoxicity because glutamate receptor blockers like MK-801 and CNQX increased significantly neuronal survival. MK-801 and CNQX, which block NMDA and AMPA/Kainate receptors, respectively, had additive effects. Expression of AMPA/Kainate glutamate receptors in mixed adult retinal cell cultures was attested by patch clamp recording. In newborn rat retinal culture, glutamate excitotoxicity was significantly reduced by addition of the L-type Ca(2+) channel blocker, diltiazem. In in vivo experiments, the increase in ocular pressure induced a decrease in cell number in the inner nuclear and ganglion cell layers. When animals received diltiazem injections, the ischemic treatment induced a less severe reduction in retinal cells; this neuroprotection was statistically significant in the ganglion cell layer.

CONCLUSION

These results are consistent with previous studies suggesting that Ca(2+) channel activation contributes to retinal cell death following either glutamate excitotoxicity or retinal ischemia. Under both conditions, the L-type Ca(2+) channel blocker, diltiazem, can limit cell death. These results extend the potential application of diltiazem in retinal neuroprotection to retinal pathologies involving glutamate excitotoxicity and ischemia.

Advanced glycation end products can induce glial reaction and neuronal degeneration in retinal explants

BACKGROUND/AIMS

Neuronal degeneration has been reported to occur in diabetic retinopathy before the onset of detectable microvascular abnormalities. To investigate whether advanced glycation end products (AGE) could be directly responsible for retinal neurodegeneration, retinal explants were incubated with glycated bovine serum albumin (BSA).

METHODS

Retinal explants obtained from non-diabetic adult rats were incubated 4 days with or without 200 mug/ml glycated BSA. Neural apoptosis was quantified by terminal dUTP nick end labelling (TUNEL) binding and immunostaining with anti-cleaved caspase-3 antibody. Expression of glial fibrillary acidic protein (GFAP) was localised by immunofluorescence.

RESULTS

TUNEL and cleaved caspase-3 positive cells increased significantly by 2.2-fold and 2.5-fold in retinal explants incubated in glycated BSA (p<0.05), respectively. The ganglion cell layer was the most sensitive retinal layer to the glycated BSA. Neuronal degeneration was confirmed by the increased GFAP labelling in Müller glial cells from retinal explants treated with glycated BSA.

CONCLUSION

These results suggest that AGE could induce retinal neurodegeneration in the absence of blood perfusion. Cells in the ganglion cell layer appeared to be the most sensitive as in diabetic retinopathy and its animal models. AGE toxicity could therefore contribute to the early pathological mechanisms of diabetic retinopathy.

Excessive activation of cyclic nucleotide-gated channels contributes to neuronal degeneration of photoreceptors

In different animal models, photoreceptor degeneration was correlated to an abnormal increase in cGMP concentration. The cGMP-induced photoreceptor toxicity was demonstrated by applying the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine on retinal explants. To assess the role of cGMP-gated channels in this cGMP toxicity, the Ca(2+) channel blockers verapamil and L- and D-diltiazem, which block cGMP-gated channels with different efficacies, were applied to in vitro animal models of photoreceptor degeneration. These models included: (i) adult rat retinal explants incubated with zaprinast, a more specific inhibitor of the rod phosphodiesterase than 3-isobutyl-1-methylxanthine and (ii) rd mouse retinal explants. Photoreceptor apoptosis was assessed by terminal dUTP nick end labelling and caspase 3 activation. Effects of the blockers on the synaptic rod Ca(2+) channels were measured by patch-clamp recording. In the zaprinast-induced photoreceptor degeneration model, both diltiazem isomers rescued photoreceptors whereas verapamil had no influence. Their neuroprotective efficacy was correlated to their inhibition of cGMP-gated channels (l-diltiazem>d-diltiazem>verapamil=0). In contrast, all three Ca(2+) channel blockers suppressed rod Ca(2+) channel currents similarly. This suppression of the currents by the diltiazem isomers was very weak (16.5%) at the neuroprotective concentration (10 microm). In rd retinal explants, both diltiazem isomers also slowed down rod degeneration in contrast to verapamil. L-diltiazem exhibited this effect at concentrations ranging from 1 to 20 microm. This study further supports the photoreceptor neuroprotection by diltiazem particularly in the rd mouse retina, whereas the absence of neuroprotection by verapamil further suggests the role of cGMP-gated channel activation in the induction of photoreceptor degeneration.

Retinal-cell-conditioned medium prevents TNF-alpha-induced apoptosis of purified ganglion cells

PURPOSE

Retinal ischemic processes occurring in glaucoma or diabetic retinopathy induce the secretion of tumor necrosis factor (TNF)-alpha. This cytokine was reported to be either toxic to or protective of retinal ganglion cells (RGCs). In the present study, its effect on RGCs was analyzed in different culture conditions.

METHODS

Adult rat RGCs were prepared in mixed retinal cell cultures and in purified cultures. They were incubated in normoxic or ischemic conditions, in the presence or absence of TNFalpha and/or conditioned media isolated from rat retinal glial cell cultures and from adult mixed retinal cell cultures.

RESULTS

In mixed retinal cell culture, RGCs were insensitive to TNF-alpha, whereas it induced their degeneration in purified adult RGC cultures. This TNFalpha-elicited toxicity was suppressed by TNFalpha-R1-neutralizing antibodies or caspase 8/10 inhibitors. Analyses of mRNA and protein content in purified RGCs revealed a time-dependent reduction in the expression of the inhibitor of caspase-8, c-FLIP. c-FLIP mRNA was also undetectable after 5 days of culture in the presence of TNFalpha. The retinal cell-conditioned medium protected the RGCs from TNFalpha-induced death and prevented the decrease in c-FLIP mRNA and protein in purified cultures. This medium promoted NF-kappaB translocation in purified RGCs, whereas an NF-kappaB inhibitor induced RGC death in mixed retinal cells.

CONCLUSIONS

The results confirm that TNFalpha can induce RGC death by TNF-R1 activation. They indicate, however, that other retinal cells can release a molecule that promotes NF-kappaB translocation in RGCs, the synthesis of the anti-caspase-8, c-FLIP, and thereby prevents TNFalpha-mediated RGC death.

Purification of mammalian cone photoreceptors by lectin panning and the enhancement of their survival in glia-conditioned medium

PURPOSE

In retinal diseases characterized by photoreceptor degeneration, the main cause of clinically significant vision loss is cone, rather than rod, loss. In the present study, a technique was designed to purify cones to make it possible to screen for neuroprotective molecules.

METHODS

A suspension of porcine retinal cells was incubated on coverslips coated with the peanut agglutinin (PNA) lectin, which selectively binds to cones. Cones were identified and quantified by using an antibody specific for cone arrestin. Their identity and viability were also assessed by single-cell RT-PCR and patch-clamp recording.

RESULTS

This panning method provided a population of cones that was 80% to 92% pure, depending on the counting strategy used. The panned cells contained both short (S)- and medium/long (M/L)-wavelength opsin cones. The panned retinal cells exhibited the physiological signature of cone photoreceptors and single-cell reverse transcriptase-polymerase chain reaction (RT-PCR) showed that they expressed the cone arrestin mRNA. Most (69%) cone photoreceptors produced neurites and survived for up to 7 days when cultured in a glia-conditioned medium, whereas very few (4%) survived after 7 days in the control medium.

CONCLUSIONS

This PNA-lectin-panning method can provide highly pure and viable mammalian cones, the survival of which can be prolonged by glia-conditioned medium. Because PNA lectin binds to cone photoreceptors from various species in both normal and pathologic conditions, this technique should enable the screening of neuroprotective molecules like those released by glial cells and enable the physiological, genomic, and proteomic characterization of cones.

Analysis of Euglena gracilis alpha-, beta- and gamma-tubulin genes: introns and pre-mRNA maturation

We have cloned and analyzed alpha-, beta- and gamma-tubulin genes from Euglena gracilis. The gamma-tubulin genes are 6-10 times longer than the alpha- and beta-tubulin genes, owing to the presence of numerous introns. These introns are all of the conventional type, whereas the alpha- and beta-tubulin genes contain both conventional and non-conventional introns. This is the first time that both types of introns have been found in the same gene. In the E. gracilis genome there are two genes for each tubulin, but the level of gamma-tubulin mRNA is 60 times lower than that of alpha- and beta-tubulin RNAs. The distinctive structure of gamma-tubulin genes prompted us to investigate the maturation of its pre-mRNA. We show that trans-splicing occurs before the cis-splicing of the first intron of the pre-mRNA and that polyadenylation occurs after the cis-splicing of the last intron of the pre-mRNA. We propose that mRNA processing is likely to play a role in regulating the amounts of different tubulins in E. gracilis.

The cox1 gene from Euglena gracilis: a protist mitochondrial gene without introns and genetic code modifications

We present the nucleotide sequence of the cox 1 gene encoding subunit 1 of cytochrome c oxidase in Euglena gracilis, the first report on a mitochondrial gene from this protist. Its study reveals that the Euglena mitochondrial genome does not appear as a compact and homogeneous structure and that its A+T content is high (about 76%) whereas this value is less than 50% in nuclear DNA. The Euglena cox1 gene does not exhibit any intron, and an amino-acid alignment of Euglena COX1 with homologous proteins shows that the universal genetic code is used. Comparisons of the genomic and cDNA sequences of Euglena cox1 indicate that the transcript does not undergo RNA editing as found in trypanosomes and in higher plants. The phylogeny obtained with COX1 protein sequences is in agreement with that obtained with nuclear rRNA sequences and places Euglena and Trypanosoma far apart from other eukaryotes. This result strengthens the hypothesis that these protists represent the earliest mitochondrion-containing organisms.

Structure and expression of Euglena gracilis nuclear rbcS genes encoding the small subunits of the ribulose 1,5-bisphosphate carboxylase/oxygenase: a novel splicing process for unusual intervening sequences?

In the protist Euglena gracilis, the small subunit of the chloroplast enzyme ribulose 1,5-bisphosphate carboxylase/oxygenase is encoded by nuclear rbcS genes and synthesized as a polyprotein precursor containing eight mature small subunit molecules. This large precursor is encoded by at least eight different nuclear genes as ascertained by transcript analysis. The structure of three rbcS genes was established and the coding sequences were found to be interrupted by many intervening sequences (IVS). Apart from the first 5' intron involved in trans-splicing, none of these IVSs obeys the GT-AG rule characteristic of introns in higher eukaryote genes. Surprisingly, these IVSs are located at identical positions within the three genes studied. Moreover, extensive sequence homologies were found between IVSs located in the same gene as well as in different genes. The sequences of these homologous IVSs differ only by inserted and/or deleted sequences. The striking conservation of the 5' and 3' regions of these IVSs is correlated to their potential secondary structures. These structures, which bring the IVS extremities together with the exon boundaries, could be involved in a novel splicing process. The second 5' IVS is shown to be excised before the addition of the spliced leader sequence to the pre-mRNA. Similarly, two 3' IVSs are excised before the polyadenylation step. These results suggest that IVS splicing is faster than eukaryotic genomic cis-splicing and involves components other than those of the classical spliceosomes.

The Euglena gracilis rbcS gene contains introns with unusual borders

We have recently shown that, in Euglena gracilis, leader sequences are transferred by trans-splicing to the vast majority of cytoplasmic mRNAs. Trans-splicing is involved in the maturation of the rbcS transcript, which encodes eight small subunits of the ribulose 1,5 bisphosphate carboxylase/oxygenase. In this report, we show that the Euglena rbcS gene introns are different from introns found in plant rbcS genes. In addition these introns do not have the conserved 5' and 3' border sequences found in introns of eucaryotic nuclear-encoded pre-mRNAs, and they do not present any homology with self-splicing introns of groups I and II. Secondary structure analyses show that the 5' and 3' ends of Euglena introns can base-pair, suggesting that an unusual splicing mechanism exists in Euglena.

In Euglena, spliced-leader RNA (SL-RNA) and 5S rRNA genes are tandemly repeated

In Euglena gracilis, a 26 nucleotide leader sequence (spliced leader sequence = SL) is transferred by trans-splicing to the 5' end of a vast majority of cytoplasmic mRNAs (8). The SL originates from the 5' extremity of a family of closely related snRNAs (SL-RNAs) which are about 100 nucleotide long. In this paper we present the nucleotide sequences of two SL-RNA genes, confirming the sequences previously established by sequencing purified SL-RNAs. Although some SL-RNA genes are dispersed throughout the genome, we show that the majority of SL-RNA genes are located on 0.6 kb repeated units which also encode the cytoplasmic 5S rRNA. We estimate that the copy number of these repeated units is about 300 per haploid genome. The association of SL-RNA and 5S rRNA genes in tandemly repeated units is also found in nematodes but paradoxically does not exist in trypanosomes which are phylogenically much closer to Euglena. We also show that a high number of sequences analogous to the 26 nucleotide SL are dispersed throughout the genome and are not associated with SL-RNAs.

Short leader sequences may be transferred from small RNAs to pre-mature mRNAs by trans-splicing in Euglena

Very closely related short sequences are present at the 5' end of cytoplasmic mRNAs in Euglena as evidenced by comparison of cDNA sequences and hybrid-arrested translation experiments. By cloning Euglena gracilis nuclear DNA and isolating the rbcS gene (encoding the small subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase), we have shown that the short leader sequence does not flank the nuclear gene sequence. The leader sequences were found to constitute the 5' extremities of a family of small RNAs. Sequencing six members of this family revealed a striking similarity to vertebrate U snRNAs. We propose that a trans-splicing mechanism transfers the spliced leader (SL) sequence from these small RNAs (SL RNAs) to pre-mature mRNAs. Transfer of leader sequences to mRNAs by trans-splicing has been shown only in trypanosomes where cis-splicing is unknown, and in nematodes where not more than 10% of the mRNAs have leader sequences. Our results strongly suggest that Euglena is a unique organism in which both a widespread trans-splicing and a cis-splicing mechanism co-exist.

RNA structural elements for expression in Escherichia coli. Alpha 1-antitrypsin synthesis using translation control elements based on the cII ribosome-binding site of phage lambda

Analysis of a series of lambda cII::alpha 1-antitrypsin (alpha 1AT) gene fusions of different sizes showed that increased alpha 1AT expression correlated with the stabilisation of a particular computer-predicted RNA secondary structure. Moreover, significant synthesis of unfused alpha 1AT was achieved by reconstruction of this conformation to permit interaction between the upstream region of the ribosome-binding site and the first part of the alpha 1AT coding sequence. This high-level expression was dependent upon certain silent point mutations in the coding sequence, indicating that RNA primary and secondary structure determinants can operate in concert to dictate the efficiency of protein synthesis.

Altered specificities of genetically engineered alpha 1 antitrypsin variants

Seven active site variants of human alpha 1-antitrypsin (alpha 1AT) were produced in Escherichia coli following site-specific mutagenesis of the alpha 1AT complementary DNA. alpha 1AT (Ala358), alpha 1AT (Ile358) and alpha 1AT (Val358) were efficient inhibitors of both neutrophil and pancreatic elastases, but not of cathepsin G. alpha 1AT (Ala356, Val358) and alpha 1AT (Phe358) specifically inhibited pancreatic elastase and cathepsin G respectively. The most potent inhibitor of neutrophil elastase was alpha 1AT (Leu358), which also proved to be effective against cathepsin G. The alpha 1AT (Arg358) variant inactivated thrombin with kinetics similar to antithrombin III in the presence of heparin. Electrophoretic analysis showed that SDS-stable high mol. wt complexes were formed between the mutant inhibitors and the cognate proteases in each case. These data indicate that effective inhibition occurs when the alpha 1AT P1 residue (position 358) corresponds to the primary specificity of the target protease. Moreover, alteration of the P3 residue (position 356) can further modify the reactivity of the inhibitor. Two of the variants have therapeutic potential: alpha 1AT (Leu358) may be more useful than plasma alpha 1AT in the treatment of destructive lung disorders and alpha 1AT (Arg358) could be effective in the control of thrombosis.

Antiprotease targeting: altered specificity of alpha 1-antitrypsin by amino acid replacement at the reactive centre

Alpha-1 antitrypsin (alpha 1AT) is an efficient inhibitor of the human neutrophil proteases, elastase and cathepsin G. The reactive centre P1 residue (Met358) of alpha 1AT is important in defining the specificity of inhibition; furthermore, oxidation of this residue results in a loss of inhibitor activity. There is evidence that oxidative inactivation of alpha 1AT may be involved in the pathogenesis of pulmonary emphysema associated with cigarette smoking. We have studied the effect of a series of amino acid replacements at the active centre on the inhibition properties of alpha 1AT. The mutant proteins were produced in E. coli following in vitro mutagenesis of the alpha 1AT cDNA. Alpha-1-AT (Ile358), (Ala358) and (Val358) were efficient inhibitors of both neutrophil and pancreatic elastase, but not cathepsin G. Alpha-1-AT (Ala356, Val358) and alpha 1AT (Phe358) were specific for pancreatic elastase and cathepsin G respectively. Alpha-1-AT (Leu358) inhibited both neutrophil elastase and cathepsin G. These data show that, for effective inhibition, a potential cleavage site for the protease must be displayed at the alpha 1AT active centre. In each case, replacement of Met358 led to resistance to oxidative inactivation. Since alpha 1AT (Leu358) inhibits both neutrophil proteases and is resistant to oxidation, this variant may be of increased potential for the therapy of destructive lung disorders.

Evaluation of recombinant DNA-directed E.coli produced alpha 1-antitrypsin as an anti-neutrophil elastase for potential use as replacement therapy of alpha 1-antitrypsin deficiency

alpha 1-antitrypsin (alpha 1AT) deficiency is an inherited disorder almost always associated with the development of panacinar emphysema in the fourth to fifth decades. One source of alpha 1AT for chronic replacement therapy of such individuals is that produced by E.coli directed by a cDNA coding for the human alpha 1AT molecule. Using TG1(E.coli), an alpha 1AT molecule produced by E.coli transformed with the plasmid-expressing vector pTG922, the present study shows that recombinant DNA-directed E.coli-produced alpha 1AT is as an effective inhibitor of neutrophil elastase as alpha 1AT purified from plasma. Importantly, TG1(E.coli) inhibited human neutrophil elastase with an association rate constant of 1.3 +/- 0.4X10(7) M-1 sec-1, similar to that of normal plasma alpha 1AT (1.1 +/- 0.1, p greater than 0.2). Furthermore, when TG1(E.coli) was added to alpha 1AT-deficient plasma obtained from homozygous alpha 1AT type Z individuals, the TG1(E.coli) remained functional and augmented the anti-neutrophil elastase activity of the serum proportional to the amount of TG1(E.coli) added. These observations suggest that if sufficient amounts of recombinant DNA methodology-produced alpha 1AT molecules could be safely delivered to the alveolar structures of alpha 1AT-deficient individuals, they would function to protect the alveolar walls from elastolytic attack.

Use of synthetic oligonucleotides in gene isolation and manipulation

Great progress has occurred in the techniques of synthesis of DNA molecules of defined sequences in terms of speed, length of the obtained oligonucleotides, and automation of the processes. Corresponding progress also occurred in the ways of using synthetic DNA in molecular biology and recombinant DNA research. Screening of cloned DNA sequence banks with long, unique oligonucleotides, provided a new approach to isolate the genes for proteins which are present in very small quantity. This technique can present considerable advantages over the more classical use of mixtures of oligonucleotides, in reducing the number of potentially positive clones on a primary screen, and enabling cloning with a minimum of amino acid sequence data. Synthetic oligonucleotides also provide the basis of a set of techniques for site-directed mutagenesis of DNA sequences. This allows the possibility of engineering the structure of particular proteins, and the properties of new variants can be tested by expressing the protein in a heterologous host. An example of this approach is the production of variants of human alpha 1-antitrypsin. A variant where valine replaces the methionine at the active site is equally active as an antielastase, but no longer susceptible to oxidative inactivation. A second variant, where arginine replaces the methionine, now functions as an antithrombin, but no longer inhibits elastase. Total gene synthesis is now feasible for larger and larger genes, and some of the recent strategies of whole gene synthesis are presented.

Synthesis in E. coli of alpha 1-antitrypsin variants of therapeutic potential for emphysema and thrombosis

The primary function of alpha 1-antitrypsin (alpha 1-AT), an antiprotease produced by the liver, is the inhibition of neutrophil elastase, a protease capable of hydrolysing most connective tissue components. The importance of alpha 1-AT is demonstrated by the high incidence of early-onset emphysema in individuals with hereditary alpha 1-AT deficiency (Type PiZZ), in whom serum levels of alpha 1-AT are 10-20% of normal. Oxidants in tobacco smoke can inactivate alpha 1-AT in vitro, and studies have shown that alpha 1-AT from the lungs of individuals who smoke cigarettes may also be partially inactivated, perhaps explaining the high incidence of emphysema associated with cigarette smoking. Oxidative inactivation is probably due to modification of the Met residue (Met358) at the P1 subsite position of the elastase binding site of the protein. To study the possibility of modulating the biological properties of alpha 1-AT, we have introduced selected sequence modifications at the reactive site by in vitro mutation of a cloned alpha 1-AT complementary DNA. We describe here the characterization of two alpha 1-AT analogues produced in Escherichia coli. The first, alpha 1-AT(Met385----Val), is not only fully active as an elastase inhibitor but is also resistant to oxidative inactivation. The other, alpha 1-AT(Met358----Arg), no longer inhibits elastase but is an efficient thrombin inhibitor. The active site of the latter is identical to that of the alpha 1-AT (Pittsburgh) variant, which was associated with a fatal bleeding disorder.

The influence of mRNA primary and secondary structure on human IFN-gamma gene expression in E. coli

Parameters influencing the efficiency of expression of the human immune interferon (IFN-gamma) gene in E. coli were studied by comparing a series of eight in vitro-derived gene variants. These contained all possible combinations of silent mutations in the first three codons of the mature IFN-gamma polypeptide coding sequence. Expression levels varied up to 50-fold among the different constructions. Comparison of messenger RNA secondary structure models for each variant suggested that the presence of stem-loop structures blocking the translation initiation signals could drastically decrease the efficiency of IFN-gamma synthesis. With variants displaying no stable mRNA secondary structure in the region, a C----U transition at position +11 after the AUG resulted in a 5-fold increase in expression indicating that RNA primary structure also plays an important role in expression. In addition we demonstrate that, in this system, a spacing of 8 nucleotides between the Shine-Dalgarno region and AUG was optimal for gene expression and that the steady-state production level of IFN-gamma rose exponentially with increasing rate of synthesis.

High-level production of biologically active human alpha 1-antitrypsin in Escherichia coli

A cDNA clone containing the complete human alpha 1-antitrypsin sequence was isolated from a human liver cDNA bank by screening with a chemically synthesized oligonucleotide probe. DNA sequences encoding the alpha 1-antitrypsin mature polypeptide were inserted into an Escherichia coli expression vector that allows transcription from the efficient leftward promoter of bacteriophage lambda (PL) and initiation of translation at the lambda cII gene ribosome-binding site. This construction resulted in the induction of a 45-kilodalton protein at a level of approximately 15% of total cell protein. The polypeptide produced was recognized by antisera raised against human alpha 1-antitrypsin protein and displayed normal biological activity in an in vitro antielastase assay.