Alternative DNA conformations are of particular interest as potential signals to mark important sites on the genome, contrasting with the rather uniform structure of the classical B-form DNA double helix which shows little variation as a function of its nucleotide sequence. The structural variability of CA microsatellites is particularly pronounced. These are repetitive poly (CA) · poly (TG) DNA sequences spread in all eukaryotic genomes as tracts of up to 60 base pairs long, sometimes longer, which are particularly abundant in the human genome where they are present in about 10⁵ copies [1]. Because of this abundance, combined with their frequent length variability between different individuals, they have served as a very useful source of markers in human genetics. Many in vitro studies have shown that the structure of poly (CA) · poly (TG) can vary markedly from the classical right-handed DNA double helix and adopt diverse conformations (for a review see e.g. [2] and references therein), the best known of which being left-handed Z-DNA [3,4]. In the course of our work with DNA fragments containing this repetitive sequence we have observed the formation of several alternative structures which appeared as retarded bands upon gel electrophoresis. While some of them have been shown to correspond to multistranded complexes [5,6,7], a series of closely spaced bands initially named 'bands X' [5] were drawn to our attention for two reasons. First, they migrated near the regular double-stranded form of the fragment, suggesting that they might correspond to double-stranded, not multi-stranded, structures. Second, they were bound with high affinity by proteins HMG1 and HMG2, two abundant non-histone nuclear proteins for which no double-stranded DNA substrate with such a high affinity was known. Here we describe a mechanism of formation of these structures, and their characterization as DNA loops at the base of which the DNA duplexes form a unique knot in which one of the strands of one duplex passes between the strands of the other duplex, and reciprocally, to form a semicatenated DNA junction, also called a DNA hemicatenane.

A 120 base pair (bp) DNA fragment containing a 60 bp tract of poly(CA) · poly(TG), when incubated under specific conditions and analyzed by polyacrylamide gel electrophoresis, can give rise to a series of bands. While some of these bands have been shown to correspond to multistranded complexes [5,6,7], a series of closely spaced bands initially named 'Form X' [5] were drawn to our attention for two reasons. First, they migrated near the regular double-stranded form of the fragment, suggesting that they might correspond to double-stranded, not multi-stranded, structures. Second, they were bound with high affinity by proteins HMG1 and HMG2, two abundant non-histone nuclear proteins for which no double-stranded DNA substrate with such a high affinity was known.

These bands are shown in Figure 1a, where they are noted X and migrate on this polyacrylamide gel between the two single strands of the DNA fragment. These structures are stable and can be purified by electroelution from a polyacrylamide gel ([5]; Fig. 1b lane 1). The study of their thermal stability (Fig. 1b) shows that they dissociate upon moderate heating to give the regular double-stranded fragment, which dissociates in turn at higher temperature to give its two single strands. Form X, therefore, contains both strands of the fragment in equimolar amount. Figure 1c shows the affinity of proteins HMG1 and HMG2 for Form X DNA. Under conditions where neither double-stranded nor single-stranded DNA is bound, Form X is entirely complexed with HMG1/2 even at a ratio of competitor DNA to Form X higher than 10⁶. A detailed analysis of the relative affinities of HMG1/2 to different DNA substrates (double-stranded or single-stranded DNA, loops, minicircles, cruciform, Form X) will be presented elsewhere (C.G. and F.S., in preparation).

To study the detailed structure of Form X it was necessary to purify it in sufficient amounts. We have never been able to induce its formation directly, even by incubating the double-stranded fragment in the presence of a wide variety of agents or by varying the pH between 6.0 and 9.0. But, after it was shown that the formation of multistranded structures by fragments containing a tract of poly(CA) · poly(TG) was correlated to some extent with the opening of the double helix [6,7,8,9], we found that the most efficient way of producing Form X was to dissociate the DNA strands by thermal denaturation and to let them reassociate in the presence of protein HMG1 or HMG2. In this manner, we obtain the complexes of Form X with HMG1/2, which can then be dissociated by SDS and Form X purified (Fig. 1d, lane 2). In the absence of HMG1/2, this same process gives no Form X or hardly detectable amounts of Form X.

To study the structure of Form X, experiments were performed to study its sensitivity to single-strand specific nucleases (S1 and P1 nucleases), and to chemicals which react specifically with non-B regions of DNA (diethylpyrocarbonate, hydroxylamine, permanganate). Such experiments (not shown) clearly showed a change of conformation limited to the repetitive region of the fragments, but did not allow us to determine its exact structure. For example, the hypothesis that Form X might correspond to four-stranded structures could not be ruled out at that stage. In addition, structures containing staggered single-stranded loops resulting from shifted reassociation in the repetitive region had to be considered [7, 8]. Starting with this hypothesis, in an attempt to measure the length of such loops, we set out to determine the linking number of DNA in Form X, i.e. the number of times one strand turns around the other strand in such a structure. Indeed, if Form X contains unpaired regions, the linking number is expected to decrease by one unit for each unpaired turn of double helix (10.5 bp).

To test this hypothesis, Form X obtained with a 258 bp fragment containing the same 60 bp tract of poly(CA) · poly(TG) as above was circularized, and the products obtained were compared to a series of marker topoisomers obtained by circularization of the regular linear fragment in the presence of increasing amounts of ethidium bromide [10]. To clearly resolve all the topoisomers, the analysis was performed on polyacrylamide gels in the absence or in the presence of chloroquine [11] and is shown in Figure 2. It is observed that circularized Form X (noted X_c) does not migrate like any of the marker topoisomers. It can also be seen that, unlike linear Form X (X_L), circular Form X is extremely stable and is not modified by incubation at 100°C (nor by alkaline pH, result not shown). Neither is it modified by incubation with calf thymus topoisomerase I or by human topoisomerase II, suggesting that it contains no superhelical stress. Since circular Form X does not correspond to any band in the series of topoisomer markers, we then considered that Form X might contain a large number of negative supercoils, larger than in the most supercoiled of the marker topoisomers, and that the corresponding torsional stress was absorbed and constrained by a change of conformation strictly limited to the poly (CA) · poly (TG) region and stabilized by supercoiling of the small circles. For example, a change of conformation from B-DNA to Z-DNA might have corresponded to this hypothesis.

This was shown not to be the case when Form X was inserted in a large DNA fragment, leading to the formation of 2384 bp circles in which the linking number was measured by two-dimensional agarose gel electrophoresis, the first dimension in the absence of chloroquine and the second dimension in the presence of chloroquine [11] (Figure 3). No difference of migration is visible between circles containing Form X and circles containing the regular form of the fragment. If the circles containing Form X are preincubated at 100°C before electrophoresis, exactly the same distribution of topoisomers is obtained (not shown). Therefore the presence of Form X on a DNA circle of that size does not modify its electrophoretic mobility, and it is impossible to show any change of linking number in Form X relative to regular DNA. In addition, as observed with 258 bp minicircles (Fig. 2), recutting the 2384 bp circles containing Form X shows that Form X is completely stable and resistant to heating at 100°C when contained in a covalently closed circle (Fig. 3d, lanes 3 and 4). Therefore the hypothesis of a global change of conformation of poly(CA) · poly(TG) induced or stabilized by supercoiling had to be ruled out. This experiment also shows with no ambiguity that Form X contains two ends only, since ligation of Form X with the vector yields almost exclusively monomeric circles, even in the presence of an excess of vector. Therefore Form X can only be a two-stranded structure.

At this stage of the work, Form X looked as a paradox: a stable non-B double stranded DNA structure, with no visible change of DNA linking number. This puzzle was resolved by ligating hairpin oligonucleotides at the ends of linear Form X. Figure 4 shows that Form X remains stable in molecules with closed ends, and resists heating to 100°C and treatment by alkaline pH (lanes 4 and 5), as do circular molecules, but unlike open linear molecules. It should also be noted that upon adding a hairpin oligonucleotide at one end only, Form X is not as stable and can be dissociated by heat treatment (lane 7), although not as easily as when it is contained in an open linear fragment.

These results show that both DNA strands in Form X do not simply run side by side as in regular DNA, but that they are somehow associated in a knot. A simple model appears if one considers the process used for producing Form X. During their reassociation, the two repetitive DNA strands do not necessarily pair in perfect register, but can also pair with a shift in the repetitive poly (CA) · poly (TG) region. In such a case, one should expect a pause when the reassociation process reaches the sides of the repetitive region. We suggest that during this pause one of the single strands of one end can insert in the fork formed by the two single strands at the opposite end, possibly through interactions between the short single-stranded repetitive sequences remaining on both sides of the central double-stranded region (Fig. 5). Then, complete pairing of the non repetitive sequences at both ends yields the formation of a loop at the base of which two duplexes cross, with one of the strands of one duplex passing between the two strands of the other duplex, and reciprocally, to form a structure which is schematically represented in Figure 5. Several parameters of such a structure can vary, including the location of the junction within the repetitive nucleotide sequence, the size of the loop, and the DNA linking number inside the loop. This is probably the explanation for the number of different bands shown by Form X on a polyacrylamide gel, where up to seven bands can be seen depending on the gel concentration.

The role of HMG1/2 in the process of formation of Form X can be envisioned at two stages. On one hand, HMG1/2 increases the flexibility of DNA [12, 13] which should facilitate the formation of a loop in the central region. On the other hand, the known affinity of HMG1/2 for DNA junctions should help stabilize the transient junction before complete pairing of the non-repetitive terminal regions.

The possibility that two DNA duplexes might associate to form such junctions has been proposed previously [14,15,16,17,18,19], and such structures have been termed 'hemicatenanes' to reflect the fact that each duplex is linked to the other duplex by only one of its strands, although from an etymological point of view the term 'semicatenane' formed with two Latin roots would seem more correct. We have continued to use the term 'Form X' which we used before we began to determine its structure. Semicatenanes have been considered to explain the association of meiotic chromosomes [16, 17], although a double Holliday junction structure has so far been preferred. Semicatenanes have also been proposed as intermediates in the replication of the genome of SV40 virus, to explain some intermediates which seem not to correspond to fully catenated structures [15, 18]. The possibility for producing such structures in vitro and their remarkable stability should now allow the study of their characteristics, which should, in turn, facilitate the investigation of their significance in vivo. It should be noted that the repetitive sequence is required only because of the particular process used to prepare Form X, but that there should be no theoretical objection to the existence of such structures with non-repetitive sequences.

The remarkable stability of this structure makes it possible to consider several experiments. For example, it should be possible to insert such structures into vectors and to introduce them into living cells, allowing to study their evolution and their effect on the biological activity of the DNA molecule in which they are inserted. Equally interesting would be the study of the enzymes and more generally of the proteins able to interact with these structures. The fact that proteins HMG1 and HMG2, two of the most abundant non-histone proteins, bind to Form X with very high affinity (C.G. and F.S., in preparation) already suggests that this structure might play a role in the function of the genome. In addition, the formation of DNA loops has often been proposed to explain several chromosomal structures or many regulation processes (see e.g. [20,21]), and the semicatenated loop described here is certainly among the most stable of all DNA loops observed so far.

An alternative DNA structure named Form X, which was observed previously by polyacrylamide gel electrophoresis of DNA fragments containing a tract of the CA microsatellite poly (CA) · poly (TG) but had not been characterized, has now been identified as a DNA loop maintained at its base by a semicatenated DNA junction (Fig. 6). Structures containing DNA hemicatenanes had been previously suggested to exist in the cell but had not been isolated before. The possibility to prepare such structures, combined with their remarkable stability, should allow one to study their evolution and their possible function when introduced into living cells.

Vaccinia virus is a large DNA virus that replicates in the host cell cytoplasm in granular sites called virosomes [1]. It encodes at least two protein kinases belonging to the cellular family of serine/threonine protein kinases, the products of the B1R [2,3] and F10L genes [4]. The F10L kinase is encapsidated in the virion and plays an essential role in virion morphogenesis [5,6]. The B1R protein kinase is expressed early in infection, is found in the virosomes, and is also packaged into virions [7]. It appears to be an essential viral protein, and temperature-sensitive mutations that map to the B1R gene produce virus that cannot replicate its DNA at the restrictive temperature [2,8]. The B1R kinase does not appear to have a broad substrate specificity, and, although it has some activity against the acidic protein, casein, this is a poor substrate compared with the enzyme's known physiological substrates. Three proteins which become phosphorylated during infection of cells with vaccinia virus have been shown to be substrates of the B1R protein kinase in vitro. Two of these are the ribosomal proteins Sa and S2 [9,10] and the third is the product of the H5R open reading frame of the vaccinia virus genome [10,11].

The fact that the B1R kinase phosphorylates these proteins in vitro does not, of course, prove that it is responsible for their phosphorylation during infection by virus in vivo. However, one piece of evidence consistent with this possibility is that all these proteins have multiple phosphorylation sites predominantly involving threonine (rather than the more usual serine) residues, and it is threonine residues on the proteins that the B1R kinase phosphorylates in vitro. In the case of protein H5R - the subject of the current work - there is further reason to believe that the B1R kinase contributes to the phosphorylation in vivo. In a mutant strain of virus, temperature-sensitive for B1R, the proportion of underphosophorylated H5R decreases at the restrictive temperature (G. Beaud and R. Beaud, unpublished).

In cells infected with a temperature-sensitive mutant of the B1R gene, the proportion of underphosphorylated H5R protein decreases at the restrictive temperature, showing that the B1R protein kinase controls the phosphorylation state of the H5R protein synthesized at the early stage of vaccinia virus infection [12]. It has recently been shown that most of the H5R protein is found in virosomes, although some of the more highly phosphorylated forms of the protein appear to be cytoplasmic, suggesting multiple roles in vaccinia virus development [13]. Kovacs and Moss [14] have demonstrated that the H5R protein is, in fact, equivalent to the late stage-specific transcription factor VLTF-4, and Black et al. [15] have showed that it associates with protein G2R, a putative late transcription elongation factor. In contrast, studies of a dominant temperature-sensitive mutant of H5R by DeMasi and Traktman [16] suggest a role in virion morphogenesis. A knowledge of the phosphorylation sites on the H5R protein is needed to test whether phosphorylation has a role in either of these processes, and we have made the first steps in this direction by identifying two threonine residues in the protein that are substrates for the B1R protein kinase.

Vaccinia virus protein H5R isolated from infected HeLa cells was labelled in vitro with [γ-³²P]ATP by recombinant B1R protein kinase, all as described in Experimental. The products of the reaction were digested with V8 protease and then passed through a SEP-PAK cartridge to remove the majority of the unreacted [γ-³²P]ATP. The radioactive V8 peptides recovered from the cartridge were then applied to a reverse-phase HPLC column and eluted with a gradient of acetonitrile. This removed the residual [γ-³²P]ATP, and the majority of the radioactivity associated with peptides eluted as a doublet at 20-30% acetonitrile (designated 3 in Fig. 1a), which was collected, concentrated by centrifugal evaporation, redissolved and applied to a microbore HPLC column to resolve the peptides further. A first run using a 15-25% gradient of acetonitrile in 0.1% trifluoroacetic acid revealed that this preparation of the radioactive doublet contained perhaps a dozen peptides (Fig. 1b). As these were incompletely resolved, two regions of the gradient containing the most radioactive peaks (designated in the figure as I and II, containing gradient fractions 30-33 and 35-36, respectively) were collected and reapplied in turn to the microbore column, but this time the gradient for elution comprised 0-50% acetonitrile in 20 mM NaCl without trifluoroacetic acid. Insufficient radioactivity was recovered from the peaks resolved in fraction I for sequence analysis. Fraction II yielded more radioactive peptides, separated from adjacent peptides and partly resolved from each other (peaks 28 and 29 in Fig 1c).

The two peaks, 28 and 29, from the microbore HPLC separation of fraction II (Fig. 1c) were separately subjected to sequence analysis using an Applied Biosystems gas phase sequencer, and the radioactivity of the PTH derivative at each cycle of Edman degradation was determined by scintillation spectrometry. The results of thirteen cycles of automatic Edman degradation on peak 29 are shown in Fig 2. The first seven residues of the peptide were identified as YHQTTEK, which correspond exactly to residues 81-87 of H5R and do not occur elsewhere in the protein [17]. (The preceding residue 80 in the protein is glutamic acid, consistent with the substrate specificity of V8 protease.) The ³²P was released from the peptide predominantly at cycles 4, 5 and 6. As there is always carry-over from one cycle to the next, we can conclude that both residues Thr-84 and Thr-85 of H5R are phosphorylated by the vaccinia virus B1R protein kinase. Analysis of fraction 28 gave similar results to those for fraction 29 (not shown).

In many cases it has been found that peptides containing the phosphorylation sites of proteins are also substrates for the protein kinase catalysing the phosphorylation of the protein. We therefore synthesized the peptide RRIEEYHQTTEKN, which represents amino acid residues 78-88 of H5R preceded by two arginine residues (not present at positions 76-77 of the protein) to make the peptide easier to handle. As preliminary studies indicated that this served as a substrate for the B1R kinase we had two additional peptides constructed in one of which Ala replaced Thr-84 and in the other of which it replaced Thr-85. After incubation with B1R kinase and [γ-³²P]ATP the peptides were subjected to high-voltage thin layer electrophoresis to resolve the peptides from the unreacted [γ-³²P]ATP. Fig 3 shows that although there was some labelling of the two Ala-replacement peptides, this was much less than with the parent peptide. (Measurement indicated that this was about 10% of incorporation into the parent peptide). It would therefore appear that the recognition specificity for phosphorylation at either Thr residue includes the other Thr, despite the fact that in one case it is on the N-terminal side of the phosphorylated residue, and in the other case on the C-terminal side.

We have identified two phosphorylation sites for vaccinia virus protein kinase B1R in viral protein H5R, the first phosphorylation sites to be determined for this protein kinase. Although we have not established whether these same sites are also phosphorylated during viral infection, our work provides important biochemical information on the substrate specificity of the protein kinase. The sequence containing the two phosphorylated threonine residues - EEYHQTTEKNSP - is neither particularly basic or acidic, consistent with the failure of basic or acidic model peptides to serve as substrates for the B1R protein kinase. It is interesting that this does not resemble sequences reported to be phosphorylated by other protein kinases, including those phosphorylated by CK1, the known cellular protein kinase most closely related to the B1R kinase. This latter generally requires a phosphoserine three amino acids to the N-terminal side of the phosphate-acceptor site, although in some artificial substrates a highly acidic N-terminal region can substitute for this [18]. One interesting biochemical question is what residues in the target sequence we have identified are the determinants for the substrate-specificity for the B1R protein kinase. In one approach to this we have examined the sequences of other known substrates for the B1R kinase (ribosomal proteins Sa and S2) but have been unable to identify potential similarities to the H5R phosphopeptide. In this regard it should be mentioned that although the synthetic peptide RRIEEYHQTTEKN did serve as a substrate for the enzyme (Fig. 3), we found its phosphorylation to be weak compared with that produced by other protein kinases on their model peptides (Approximately 0.1 mol phosphate per mol of peptide, compared with 1.8-2.0 mol phosphate per mol H5R protein). Although the additional arginine residues in the peptide complicate the interpretation, it may well be that there are three-dimensional determinants - absent from the synthetic peptide - that contribute to the substrate specificity of the enzyme. There are precedents for this in some other protein kinases [18].

The vaccinia protein H5R has been shown by isoelectric focusing to exist in at least four differently charged forms [11,13], suggesting that there are at least least five different phosphorylation sites on the protein. Consistent with this are the results of experiments (not shown) in which we employed recombinant H5R in which Thr-84 and Thr-85 had been replaced by Ala residues, and found that this still served as a substrate for the B1R protein kinase. It is possible that one or more of these additional sites is present in the HPLC fraction I of Fig 1b, the peptides resolved from which were insufficiently radioactive for sequence analysis.

Although this work represents the first identification of phosphorylation sites for the vaccinia B1R kinase, it remains to be established whether these same sites are phosphorylated in vivo, and, if so, with what physiological significance. However it should now be possible to address this question by constructing recombinant virus with amino acid substitutions at the positions of phosphorylation.

Vaccinia virus protein kinase B1R phosphorylates the virus protein H5R in vitro at the threonine residues Thr-84 and Thr-85 within the region:

EEYHQTTEKNSP

A synthetic peptide based on this sequence also acted as a substrate. We conclude that this sequence determines, at least in part, the substrate specificity of the vaccinia B1R protein kinase, although it is unclear which amino acid residues are the key determinants within this sequence. There are other phosphorylation sites for the kinase on protein H5R, but these remain to be determined.

Cholinesterases (ChEs) are evolutionarily conserved type B carboxylesterase enzymes that share extensive sequence homology. In vertebrates two types of ChEs were identified based on their distinct substrate specificity and inhibitor sensitivity. The acetylcholinesterase (AChE; EC 3.1.1.7) specifically catalyses the hydrolysis of acetylcholine and is subjected to marked inhibition by its own natural substrate. In contrast, butyrylcholinesterase (BChE; EC 3.1.1.8) is capable of degrading a wider range of choline esters and is not inhibited by its substrate [1, 2]. AChE is selectively inhibited by BW 284c51, while BChE is specifically inhibited by tetraisopropylpyrophosphoramide (iso-OMPA) [3]. AChE is widely distributed in the nervous system and its role in rapidly terminating nerve impulse by hydrolysing acetylcholine in cholinergic synapses is well documented [1]. BChE is produced in the liver and enriched in the circulation. In addition, it is also present in adipose tissue, intestine, smooth muscle cells, white matter of the brain and many other tissues [4]. The exact physiological function of BChE is still elusive. It is generally viewed as a back up for the homologous AChE and to act as a scavenger for anticholinesterase compounds [5].

The presence of ChEs in tissues that are not cholinergically innervated provides the most compelling evidence for the view that AChE and BChE may have functions, other than the termination of cholinergic neurotransmission. There is considerable body of evidence to suggest that ChEs may be involved in embryonic neural development, including a role in cell proliferation, differentiation and cell adhesion [6, 7]. ChEs may also have a causative/permissive role in various pathological conditions as exemplified by the overexpression of ChE genes in various types of tumours and presence of abnormal levels of ChEs with altered properties in Alzheimer's disease [8, 9]. The histochemical staining of esterase activity on ChE developed by Koelle and Friedenwald [10] and modified by Karnovsky and Roots [11] has been extensively used to elucidate the functions of ChEs, examine their tissue specificity, developmental alterations and pathological changes from many species [2].

Other than the predominant choline esterase activity, ChEs also display a genuine aryl acylamidase (AAA) activity capable of hydrolysing the synthetic substrate o-nitroacetanilide into o-nitroaniline and acetate [12]. Apart from being strongly inhibited by choline esters and classical ChE inhibitors, this AAA activity is susceptible to selective inhibition by 5-hydroxytryptamine (serotonin) [12, 13, 14]. The characteristic feature of the AAA activity associated with human serum BChE is its several fold activation by tyramine [12]. The natural substrate or the precise physiological role of the AAA activity on ChEs is not known.

Studies on the tissue specific and developmental regulations/alterations of AAA on ChEs have not been attempted so far partly due to the lower specific activity of the AAA activity on ChEs and mainly due to the absence of a specific method to visualise this activity. In the present paper, a method to visualise the AAA activity on ChEs in polyacrylamide gels using human serum BChE and electric eel AChE as models is described.

The principle of the staining procedure is depicted in Figure 1. The AAA on ChEs acts on ONAA to release acetate and o-nitroaniline. o-nitroaniline then reacts with nitrous acid (provided by the reaction of sodium nitrite with HCl) at about 0°C yielding the corresponding diazonium compound. The reaction mixture must be kept very cold during the process, otherwise the diazonium compound may be partially hydrolysed to the corresponding phenol. The diazonium compound then complexes with NED to form purple coloured azo dye complex. This coupling reaction is an electrophilic substitution reaction [15].

The concentrations of sodium nitrite and HCl in the staining solution were found to be critical, since the generated nitrous acid is the one that reacts with o-nitroaniline to form the diazonium compound. For optimal detection of the AAA activity, a concentration of 0.1 % (w/v) sodium nitrite and 1 N HCl were found to be required. Similarly, studies with varying concentration of NED indicated that the maximum intensity of colour developed was with 0.75 % (w/v) of the colouring reagent.

The stained gel pattern for AAA on AChE/BChE using the above protocol is shown in Figure 2. It has been shown that the AAA activities on both AChE and BChE can be visualised by this method. Further, selective staining for the AAA activities on either BChE or AChE in a sample containing both these enzymes has been demonstrated by the use of the specific inhibitors of BChE and AChE, namely, tetraisopropyl pyrophosphoramide (iso-OMPA) and 1,5-bis(4 Allyldimethylammoniumphenyl) pentan-3-one dibromide (BW 248c51), respectively. In addition, the intensity of staining increased linearly with the enzyme concentration (in terms of AAA activity) applied in the gel (Figure 3). From Figure 3, it is also clear that, to visualise a recognisable AAA activity band in the gel, at least 0.07 U of the enzyme is required.

Prolonged incubation of the stained gels in the acid solution caused a rapid decrease in the intensity of the purple colour bands. Treating the gels with 0.3 M TCA for 30 min at 4°C prevents this loss in colour. The stain was further fixed in the gels by changing the pH to alkaline side with Tris-HCl buffer, pH 8.6. The colour of the bands changed from purple to brick red (Figure 4-A) upon changing the pH from acid to alkaline side. The gels can be stored for weeks in this alkaline buffer without loss in the intensity of the bands. Documentation of the stored gels, if necessary, can be done after treating the gels with 0.3 M TCA which brings back the original purple colour, but with a bluish tint (Figure 4-B).

α-Naphthol and β-naphthol were also tried as colouring agents instead of NED. In alkaline conditions, α-naphthol produced a red coloured band, while β-naphthol produced an orange coloured band (Figure not given). When compared to the colour developed with NED, the colour developed with either of the naphthols was faint and diffused. Moreover, the fixing of the stain in the gels was also difficult.

The AAA activities, other than those associated with cholinesterases (like those found in human liver, rat serum, pineal gland) and amidases that utilise o-nitroacetanilide as substrate can be visualised by the method described above. This method would also allow detection of such activities in crude tissue extracts, however, a minimum of 0.07 U of the enzyme has to be loaded per well to clearly visualise the activity in gels, which is the limit of detectability of this method.

This study describes a novel method to visualise the AAA activity on ChEs in polyacrylamide gels. The method has been shown to be sensitive and also can selectively detect either of the ChE's AAA activity. Use of this method to visualise AAA activity in tissue sections, however, needs further refinement/modifications to enhance the sensitivity. This is because, any tissue, at a particular locus might not have AAA activity to the extent of 0.07 U. Nevertheless, this study is the first successful attempt to visualise the AAA activity on cholinesterase in vitro.

The support by Defence Research & Development Establishment (DRDE), Gwalior, India (Project ID No. TC/05414/DRDE/Project-DRDE-149, dated 23 June, 1998) and the award of Senior Research Fellowship to L.J. by the Council of Scientific and Industrial Research, India, and DRDE are gratefully acknowledged.

AAA, aryl acylamidase; AChE, acetylcholinesterase; BChE, butyrlycholinesterase; BW 284c51, 1,5-bis(4-Allyldimethylammoniumphenyl) pentan-3-one dibromide; ChE, cholinesterase; iso-OMPA, tetraisopropyl pyrophosphoramide; NED, N-(1-napthyl)ethylenediamine; ONAA, o-nitrooacetanilide; TCA, trichloroaceticacid.

There have been plenty of epidemiological, clinical and experimental evidence that plasma high density lipoproteins levels are inversely correlated with the risk of atherosclerosis [1,2]. The contribution of enzymes and proteins associated with HDL to its process of generation and maturation have been extensively studied, both in vitro and in vivo.

The plasma cholesteryl ester transfer protein (CETP) modulates HDL levels and composition. It mediates the transfer of cholesteryl ester (CE) from HDL to triglyceride (TG) rich lipoproteins (LP), while TG is transferred in the opposite direction, to HDL. One way CE transfer from HDL to LDL may also occur [3]. CETP also promotes phospholipid (PL) transfer to human HDL [4,5,6]. CETP activity has been directly correlated with LDL cholesterol levels and inversely correlated with HDL cholesterol levels in human plasma [3,7].

PLTP, a specific phospholipid transfer protein, has been identified in human plasma [8,9] and in plasma of other vertebrate species [10]. It promotes the PL transfer from VLDL to HDL [11]. In addition to PL, PLTP transfers free cholesterol (FC) from PL/FC vesicles to HDL, although with a low efficiency [12]. Both, CETP and PLTP, can promote HDL remodelling. While CETP, together with hepatic lipoprotein lipase, stimulates the generation of small alpha-HDL, PLTP favours the emergence of large alpha-HDL particles [13]. Significantly higher levels of HDL-cholesterol were observed in human PLTP transgenic mice [14]. Furthermore, overexpression of human PLTP produced by recombinant adenovirus injection into mice, resulted in increased levels of prebeta-HDL, increased fractional catabolic rate and liver uptake of CE and PL from HDL [15].

After intravascular hydrolysis of TG rich LP by lipoprotein lipase (LPL), surface remnant components such as FC, PL and apoproteins may provide substrates for generation or modification of plasma HDL. Net transfer of PL and FC from chylomicrons and VLDL to HDL has previously been demonstrated in rats [16,17] and in human plasma after a fat meal [18,19] or during lipolysis [11]. The contribution of the lipolysed LP components to HDL formation has been reinforced by several studies where the activity of the enzyme lipoprotein lipase LPL was shown to correlate with HDL cholesterol levels in human plasma [20,21,22]. However, changes in the HDL-cholesterol concentration have not been observed in mice overexpressing LPL [23] or in LPL heterozygous knockout mice [24].

The metabolism of HDL in rats and mice differs significantly from that in humans. Part of the species differences observed in mice and in rats may result from their high levels of circulating lipases [25,26], lack of CETP [27] and high levels of PLTP [10]. Clee et al. [28] have shown that, in double transgenic mice overexpressing LPL and CETP, HDL cholesterol levels were significantly influenced by the LPL activity while no such correlation was observed in the absence of CETP expression.

In this study we have further evaluated the influence of plasma lipases and CETP on the free cholesterol and phospholipid transfer from triglyceride rich lipid emulsion similar to chylomicrons [29,30] to human, rat and mouse lipoproteins, in vitro and in vivo.

The influence of the intravascular lipase activity on the phospholipids (PL) and free cholesterol (FC) transfer from lipid emulsion (EM) to plasma lipoproteins was evaluated in human control plasma, in lipase-enriched post-heparin plasma and in post-heparin plasma containing a lipase inhibitor, tetrahydrolipstatin (THL) [34]. PL transfer was determined after 30 minutes of incubation because Tall et al. [8] have demonstrated that PL net transfer reaches the maximum after this incubation period. The post-heparin plasma lipoprotein lipase (LPL) and hepatic lipoprotein lipase (HL) activities were 2527 ± 1353 and 3747 ± 2116 nmol of fatty acid/ml/h, respectively. The control and post-heparin + THL plasmas had no detectable lipase activity, showing that 2 mM THL completely inhibited the activities of the lipases present in the post-heparin plasma. Table 1 shows that ¹⁴C-PL was preferentially transferred to HDL (p<0.001) while ³H-FC was equally distributed to LDL and HDL fractions under all 3 plasma conditions. The transfer of ¹⁴C-PL from EM to HDL was significantly stimulated (+60%) by the increase in the lipases' activities in the post-heparin plasma (p<0.001). This stimulation of the PL transfer was abolished in the presence of THL (p<0.01). Considering that HL also has phospholipase activity, the integrity of the phospholipid transferred to HDL was checked by thin layer chromatography. Ninety eight % of the ¹⁴C-PL was recovered in the PL band and no radioactivity was detected in the fatty acid bands. The proportion of ³H-FC and ¹⁴C-PL found in HDL (0.4) was different from that found in LDL (2.3) or in the EM (1.5), which suggests that these components of the EM were independently transferred to the LP fractions and not as a surface unit detached from the lipolysed EM.

The effect of plasma lipases on the PL transfer to HDL was also verified by incubating labeled EM with plasma from control and heparinized rats. Figure 1 shows the ¹⁴C-PL distribution in the LP fractions at zero (ice) or after 30 minutes at 37°C. In the control rat plasma (Fig. 1, upper panel) at time zero, the distribution of ¹⁴C-PL was 60, 13, 22 and 6% for VLDL, LDL, HDL and a fraction smaller than HDL (<HDL), respectively, while after 30 minutes at 37°C, the observed distribution was 22, 14, 57 and 8% for VLDL, LDL, HDL and fraction <HDL, respectively. It is interesting that PL transfer to HDL can be observed even at time zero. This may represent the combination of the spontaneous and facilitated transfer that occurs during and even after loading the reaction mixture to the FPLC column run at room temperature. Furthermore, rat plasma contains circulating lipases [25,26] that could be stimulating PL transfer in the control plasma. In the rat post-heparin plasma (Fig. 1, lower panel), the ¹⁴C-PL transfer was so much accelerated that the distributions at zero and after 30 minutes of incubation at 37°C were almost identical, that is, 12, 29.5, 32.5 and 26.5% at zero and 7, 31, 29.5 and 32.5% at 37°C for VLDL, LDL, HDL and fraction <HDL, respectively. HDL and the fraction smaller than HDL together accounted for 59 to 62% of the ¹⁴C-PL. Thus, comparing the PL moiety that remained in the emulsion fraction in the control at 37°C (Fig 1 upper, 22%) with the post-heparin plasma at 37°C (Fig 1 lower, 7%), we conclude that PL transfer to LDL and HDL fractions is highly stimulated by the enrichment of the rat plasma with the vascular lipases. We did not investigate the chemical nature of the fraction smaller than HDL but we speculate that it derives from HDL phospholipids or from the emulsion surface itself, and might be lysophospholipids/albumin or phospholipids/apo AI complexes.

In an attempt to evaluate the PL and FC transfer to HDL in a biological system, we injected the radioactive EM (³H-PL/¹⁴C -FC) into anesthetised functionally hepatectomized rats treated with saline (control) or heparin (LPL stimulated). The liver exclusion maneuver was done in order to minimize the differences in the residence time of the emulsion particles in the circulation of the control and LPL stimulated rats. The hepatectomy efficacy was verified by the amount of ³H-PL found in the livers at the end of the experiments; that was 1% and 3% of the injected dose in control and heparinized animals, respectively. There was a lower recovery of the labeled PL in the plasma of the heparinized animals (∼ 50%) than in control rats (∼ 90%). This suggests that the heparin stimulated lipolysis increased the plasma removal of PL by peripheral tissues. Table 2 shows the distribution of ³H-PL and ¹⁴C-FC from the EM into the plasma LP of control and heparinized hepatectomized rats. When intravascular lipolysis was stimulated by heparin, there was a 50% reduction of the ³H-PL content in the chylomicron (CM) + VLDL fraction (d<1.006) and a parallel increase in the LDL fraction that also includes remnant particles (d= 1.006 to 1.063). No significant difference was observed in the amount of ³H-PL found in the HDL fraction of control and heparinized hepatectomized rats. A similar result was observed in relation to the ¹⁴C-FC distribution, i.e., a reduction trend of ¹⁴C-FC in the CM+VLDL fraction (non significant) and a significant increase of ¹⁴C-FC transfer to the LDL fraction. However, in both cases, control and heparinized animals, the ratio ¹⁴C-FC/³H-PL differs markedly in the three LP fractions (Table 2). This indicates that, in vivo, the redistribution of the two emulsion surface lipids is independent one from another.

In order to verify the influence of CETP in the PL transfer to HDL, labeled emulsion was incubated with human plasma in the presence and absence of monoclonal antibody against CETP (TP2) and with plasma from CETP transgenic and non-transgenic mice. The CETP activity in these plasmas were 38 and 16% of cholesteryl ester transfer (4 h assay) respectively for human plasma without and with TP2, and 5 and 20% of cholesteryl ester transfer (2 h assay), respectively for nontransgenic and human CETP transgenic mice plasma. Table 3 shows that ¹⁴C-PL transfer from lipid emulsion to human HDL was reduced 40% (p<0.005) when CETP activity was inhibited 58%, indicating that CETP has PL transfer activity as previously reported [5]. However, the ¹⁴C-PL distribution in mice LP after incubation with the ¹⁴C-PL EM was almost identical for CETP transgenic and non-transgenic mice plasma (Fig. 2). The human CETP expressed in these mice might not efficiently interact with the endogenous mice HDL. To check this possibility an exogenous assay was performed using human HDL (240 ug cholesterol) as PL acceptor and minimal amounts of CETP transgenic and non-transgenic mice plasma (50 ul) as sources of CETP and of PLTP. Results obtained (data not shown) were identical to those of the endogenous assay (fig. 2). Therefore, human CETP in transgenic mice plasma does not contribute to PL transfer to either mice or human HDL.

Several studies, using different experimental approaches, have shown that the transfer of surface components of TG-rich lipoproteins during their intravascular metabolism is important to determine both level and chemical composition of the HDL subfractions. The present work has confirmed and extended previous observations showing that human and rat plasma lipoprotein lipases stimulate PLTP mediated PL transfer from TG-rich particles to HDL but do not influence the FC transfer process. Some of the previous studies [11,35,36,37,38] that used purified exogenous lipases in in vitro incubations with isolated LP displayed a potent stimulation of the PL transfer to HDL. In order to prepare a model that would mirror a physiological system more closely, we have used whole plasma and maximal endogenous lipases activity through heparin administration in vivo as well as in vitro. It is possible that other proteins released by heparin in the vascular bed may have played a role in stimulating PL transfer. The role of circulating lipases was confirmed through the use of a lipase inhibitor (THL), which abolished the stimulating effect of lipase-enriched plasma on the PL transfer to HDL (Tab. 1). According to these experiments it is likely that during fasting state, where no circulating lipases are detectable, all PL transfer results from the action of the lipid transfer proteins, PLTP and CETP, while in a post-prandial state, when lipases expectedly are more active, the PL transfer to HDL could be raised by 60% or more due to a greater substrate supply .

Noteworthy the in vitro PL transfer to HDL in the basal plasmas of rat (fig. 1, 82%) and mouse (fig. 2, 57%) was higher than in the human's (Tab. 1, 31%) and could be ascribed to the high levels of circulating lipases [26,39,40] and PLTP [10] found in those species. This could also explain the positive correlation between the lipoprotein lipase activity and HDL concentration in human plasma [20,21,22] but not in mice with genetically modified expression of LPL [23,24,28].

The in vivo studies showed that the EM PL transfer to HDL did not differ in control and in LPL stimulated (heparin treated) hepatectomized rats. Instead, the LDL density fraction was PL enriched in the heparinized animals. This could be explained by several and not exclusive possibilities. First, the lipolysis stimulation by heparin generates more remnants of the EM that would float in the same density range as LDL (1.006 - 1.063). Second, the rat plasma fraction smaller than HDL that appeared in the in vitro incubations with post-heparin plasma (Fig. 1) could also occur in vivo and float in the same density range as LDL (1.006 -1.063). In this regard, O'Meara et. al. [41] had shown that small HDL particles from heparinized hypertriglyceridemic subjects, identified in non-denaturing gel electrophoresis and by electron microscopy, floated after ultracentrifugation in a less dense range. Those authors had considered it an aberrant result of the ultracentrifugation technique. Third, the diminished availability of PL donor particles in the plasma of the heparinized rats (yield of 50%) as compared to control rats (yield of 90%) would be responsible for the apparent lack of stimulation of the PL transfer to the HDL fraction. Finally, other in vivo metabolic fates of PL would compete for the transfer process to plasma HDL particles in LPL stimulated animals.

The emulsion FC transfer to HDL was about 40% that of PL in both, in vitro (Tab. 1) and in vivo (Tab. 2) studies and it was not influenced by increased rate of intravascular lipolysis. These results suggest that the FC transfer is a slower, probably passive, process distinct from the facilitated PL transfer mechanism [8,9]. Also, these results challenge the possibility that new HDL particles are made from the EM surface peeling off during lipolysis because the relative PL/FC ratio was higher in HDL than in the CM+VLDL and emulsion fraction. Others also have shown that FC transfer to HDL is a slow process: FC increases in the HDL fraction only 5 to 8 h after a fat meal [19] or only after 2 h incubation of HDL with VLDL and purified rat heart LPL [36].

We have also confirmed previous studies [5] claiming that PL transfer from TG-rich particles to human HDL is facilitated by CETP, since its partial inhibition significantly reduced the PL transfer (Table 3). However, human CETP expressed in the transgenic mice plasma had no effect whatsoever on the PL transfer to mouse plasma lipoproteins or to human HDL. This lead us to admit that the mouse plasma PL transfer activity is so potent that some additional protein (CETP) with PL transfer activity would be irrelevant in an already saturated in vitro system.

In summary, the present findings indicate unequivocally the importance of the intravascular lipolytic mechanisms for the PLTP and CETP facilitated PL, but not FC, transfer process from TG-rich particles to HDL. PL enriched HDL would be more efficient in promoting FC efflux from cell membranes, hence accelerating the reverse cholesterol transport. These may provide the basis for the mechanism that accounts for the inverse correlation between HDL and TG plasma levels found in epidemiological studies in human populations as well as in several circumstances where plasma lipid levels are modified by pharmacological and dietary means.

Two-component systems are central elements of the bacterial signaling circuitry [1]. Signal transduction by these systems usually involves autophosphorylation of a histidine kinase on a conserved histidine residue and subsequent transfer of the phosphoryl group to a conserved aspartate on a receiver domain. Until recently, two-component systems had only been found in bacteria. In the past few years, genes coding for histidine kinase homologues and their corresponding receivers have also been discovered in eukaryotic organisms [for a review see 2]. Most of the corresponding eukaryotic gene products are part of a phosphoryl relay, which consists of a hybrid histidine kinase with a kinase and a receiver domain on the same polypeptide, a histidine phosphotransfer protein and a second receiver as part of a response regulator [3]. The function of eukaryotic two-component systems as histidine kinases was questionable until Posas et al. showed that the Saccharomyces cerevisiae gene product Sln1 acts as a histidine kinase in vitro and in vivo [4]. More recently, histidine kinase activity of the ethylene receptor Etr1 from Arabidopsis was demonstrated in vitro [5]. Further studies showed, however, that eukaryotic two-component systems do not function as independent pathways, but are often connected to serine/threonine- and tyrosine kinase cascades. Thus, the yeast Sln1-Ypd1-Ssk1 phosphoryl relay acts as an osmosensor, which activates a MAP-kinase cascade when cells are exposed to high osmolarity [6]. The Dictyostelium discoideum protein RegA consists of a N-terminal receiver domain and a C-terminal phosphodiesterase domain [7]. Phosphorylation of the RegA response regulator via a two-component phosphoryl relay in turn activates the RegA phosphodiesterase thereby causing a decrease in the intracellular cAMP level. Eukaryotic phytochromes, another class of histidine kinase homologues, were shown to act as light-regulated serine/threonine kinases in vitro instead of acting according to the histidine kinase paradigm [8]. These results suggest that eukaryotic two-component systems, although being homologues of bacterial histidine kinases and receivers, might show post-translational modifications found in the already established eukaryotic signal transduction systems.

In the amoeba Dictyostelium discoideum, several genes coding for histidine kinases have been described [[9-13]]. Deletions of individual histidine kinase genes cause different developmental phenotypes such as rapid aggregation, disproportioned fruiting body and stalk ratios or impaired spore formation [13]. Moreover, cells lacking the histidine kinase gene dokA are osmosensitive, i.e. the viability of these cells is decreased when exposed to high osmolarity for up to two hours [9]. Given the evidence that DokA is part of the osmotic response system of Dictyostelium, we have examined whether DokA shows kinase activity in an osmolarity-dependent manner.

In this paper, we present evidence that the histidine kinase homologue DokA is phosphorylated on a serine residue in vivo when Dictyostelium cells are exposed to a high osmolarity medium. We further demonstrate that the phosphorylation site is located in a domain homologous to bacterial histidine kinases and that mutation of the conserved histidine does not affect the serine phosphorylation of DokA.

In the past few years, a number of eukaryotic genes coding for histidine kinase homologues have been cloned and characterized. Recent work showed that eukaryotic two-component systems are integrated in signal transduction pathways involving phosphodiesterases and MAP kinases [6, 7, 18]. The results presented here confirm this idea. By labeling Dictyostelium cells with [³²P]-orthophosphate, we were able to show that overexpressed fragments of the Dictyostelium histidine kinase homologue DokA are phosphorylated in vivo on a serine residue. The phosphorylation was not present before and steadily increased after the osmotic shock reaching a plateau after 20 min. It should be noted that this phosphorylation reaction was only observed when cells were exposed to a hyperosmotic medium and not as a result of starvation or development.

The observed serine phosphorylation is in contrast to the classic two-component paradigm, which predicts only histidine and aspartate phosphorylation. However, several recent publications showed post-translational modification on serine and threonine residues, which are not in accordance with the above prediction. In particular, the stimulus-dependent serine/threonine phosphorylation of other eukaryotic histidine kinase homologues was recently demonstrated in vitro [8]. Phytochromes from oat and green algae contain histidine kinase-related domains, but lack the conserved histidine which is essential for autophosphorylation. However, they are able to autophosphorylate on serine/threonine residues when exposed to red light thereby acting as light-regulated protein kinases [8]. In contrast, the cyanobacterial phytochrome Cph1 acts as a light-regulated histidine kinase [19]. Considering these differences, oat and algal phytochromes have been designated as serine/threonine kinases with histidine kinase ancestry [8]. More recently, a sequence motif homologous to histidine kinases without the conserved histidine has been found in the adenylyl cyclase gene acrA in Dictyostelium [20].

It was previously suggested that DokA is part of the osmotic response system of Dictyostelium, as mutants lacking the dokA gene are sensitive against hyperosmotic stress [9]. Two-component systems are involved in osmoregulation in a variety of organisms [6]. In S. cerevisiae, the Sln1-Ypd1-Ssk1 phosphoryl relay acts as an osmosensor, which activates a MAP-kinase pathway when cells are exposed to hyperosmotic stress [6]. In hyperosmotically shocked mammalian cells, MAP-kinases are phosphorylated in vivo on threonine and tyrosine residues [21, 22]. In this respect, DokA resembles typical eukaryotic signal transduction proteins, as hyperosmotic stress causes serine phosphorylation of DokA. The observed osmotic stress-dependent phosphorylation supports the hypothesis that DokA is involved in an osmosensing signal transduction pathway in Dictyostelium [9, 23].

To address the question, whether DokA is phosphorylated in an autophosphorylation reaction, mutations of two conserved glycine residues in the ATP binding pocket of the catalytic domain of the kinase were introduced. The mutation was shown to completely inhibit ATP binding in histidine kinases [24]. Proteins carrying these mutations are not able to autophosphorylate on the conserved histidine residue [5]. In contrast, the serine phosphorylation of DokA is not affected by mutation of the conserved glycine residues G1205 and G1207, making an autophosphorylation reaction unlikely. Consistent with this finding is the fact that the mutation of the conserved histidine H1053 has no influence on the serine phosphorylation of DokA, indicating that no histidinyl-phosphate intermediate is necessary for this reaction. A physiologically relevant role of H1053 can, however, not be definitely precluded by this experimental approach as the DokA domains have been expressed in cells which contain a functional dokA allele. Interaction of the overexpressed peptides with wild type DokA might also be important for the detection of the osmotic stimulus as the DokA domains used in this work lack the N-terminal input domain [9]. Taken together, our results imply that DokA is a substrate for another serine/threonine kinase. To address this question, we analyzed the DokA amino acid sequence using the database PhosphoBase () to search for putative phosphorylation sites. More than 50 serine phosphorylation sites were predicted by this algorithm. None of these sites was in the receiver domain, which is consistent with our finding that the phosphorylation site of DokA is located in the kinase domain. The results of the sequence analysis indicate that a number of protein kinases may interact and subsequently phosphorylate DokA. It was previously shown that DokA regulates the activity of the RdeA-RegA pathway, thereby causing a transient intracellular cAMP signal reaching a maximum after 2 min in response to an osmotic shock [23]. In contrast, the phosphorylation reaction described here shows a 10-fold slower kinetic. We therefore speculate that the serine/threonine kinase which phosphorylates DokA might act downstream of the RdeA-RegA pathway as part of a feedback mechanism. The exact localization of the phosphorylated serine residue is currently under investigation.

We have demonstrated an osmotic stress-dependent serine phosphorylation of the eukaryotic histidine kinase homologue DokA in vivo. The phosphorylation does not depend on the conserved histidine residue, which is essential for the function of two-component systems and is not due to an autophosphorylation reaction. This confirms the idea that eukaryotic homologues of bacterial signal transduction systems might be integrated in signaling pathways involving serine/threonine kinases.

Over the years there have been numerous claims for the detection within prokaryotic and eukaryotic cells of antisense RNA sequences (AS). As recently reviewed by Kumar and Carmichael [1], many of the claims have been substantiated and in a subset of these, biological significance has been established. The origin of most of these AS sequences is transcription via DNA-directed RNA polymerization.

A much more controversial issue is whether there is AS RNA that arises via RNA-directed RNA polymerization. Of course, for RNA viruses that encode their own polymerase activity, such AS RNA arises as a natural part of the replication cycle for the viral genome. And yet for the subviral RNA agents of plants known as viroids, and for the human (subviral) agent known as hepatitis delta virus (HDV), the replication of the RNA genome involves a host RNA polymerase [2]. Evidence has accumulated but not established that the relevant polymerase activity is via redirection of a polymerase that normally uses host DNA as a template. For example, it is considered that HDV replication and that of some plant viroids might involve redirection of the host RNA polymerase II [3, 4]. Other plant viroids are considered to replicate via redirection of a chloroplast polymerase activity [5].

In contrast to such proposed redirection, studies in plants have found a non-viral host RNA-directed RNA polymerase activity [6,7,8]. After the gene for one of these was cloned and sequenced [9] it was soon found that homologs existed in virtually all plant species and also in fission yeast, zebrafish, Drosophila, and nematodes [10]. Recently, this activity has been of particular interest because of exciting findings in the area of post-transcriptional gene silencing and RNA interference. Many papers and reviews have been published recently on these topics and it is clear that the suppression mechanisms involve regions of double-stranded RNA and in some cases, RNA-directed RNA polymerase activity [11]. Two recent papers show clearly that specific silencing can be induced in mouse oocytes by complementary RNA sequences [12, 13], but apparently there is no involvement of RNA-directed RNA polymerase activity.

For the above reasons we were intrigued by the possible relevance of several specific claims for AS RNA and RNA-directed RNA polymerase activity in mammalian cells. In a series of papers by Volloch and coworkers AS RNA was reported for as many as 6 mammalian mRNAs and even RNA-directed RNA polymerase activity was proposed [14,15,16,17]. These papers focused on mouse β-globin in particular, but they also proposed AS to the mRNAs for serum amyloid A, insulin receptor, eukaryotic transcription factor eIF-4D, vasoactive intestinal polypeptide and syndecan. Most of these data were obtained by a procedure sometimes known as ligation-mediated polymerase chain reaction, that was actually pioneered by the Volloch lab [18]. However, in two studies the authors also presented Northern data [14, 15]. They first treated mice with phenylhydrazine (PHZ) to induce high levels of β-globin mRNA in the spleen, and reported that the RNA from the spleen contained low but significant amounts of an AS RNA species with a size of about 600 nucleotides, the size of the mature mRNA [14]. Also, by ligation-mediated polymerase chain reaction they reported determinations of both 5'- and 3'-ends for this AS RNA [14]. Volloch et al. speculated that the AS species they detected arose via some form of RNA-directed RNA polymerase activity, and further, they suggested that this was a mechanism for the cell to amplify globin mRNA species. Therefore, with these things in mind, we attempted to confirm and extend the most recent study of Volloch et al. [14].

As reviewed in the Introduction, it is important to know not only whether post-transcriptional gene silencing occurs in mammalian cells but also whether it can apply under natural conditions, that is, as a regulatory response to enhanced expression of a host RNA species. With this in mind we investigated whether the prior reports of AS to β-globin RNA in anemic mouse spleen could be confirmed and possibly extended. To do this we first established the same procedures in which phenylhydrazine injections were used to induce anemia in mice. We documented the production of a significant and transient splenomegaly (Fig. 1). This was associated with a transient 30-fold increase in the amount of β-globin mRNA per unit mass of spleen RNA, followed by a decrease to almost pretreatment levels (Fig. 2A and 2C).

From northern analyses and use of specific oligonucleotide probes, each exactly as reported by Volloch et al., we were unable to detect any AS globin RNA sequences (Fig. 2B). Furthermore, from the use of appropriate control hybridization standards, and quantitation using a bio-imager, we deduce that relative to the amount of sense RNA as 100%, the putative AS, if it exists, would have to be less than 0.004%. We thus can not understand how Volloch et al. reported that spleen cells contained 1-3% amounts of AS globin [14].

When we replaced the oligonucleotide probes with more sensitive RNA probes that were uniformly labeled and corresponded to the full-length mRNA (Fig. 2C-D), we did detect a species that appeared to be AS globin RNA (Fig. 2D). It was suspect that this species had an electrophoretic mobility indistinguishable from that of the detected sense RNA (Fig. 2C). However, as described in the results, we can conclude that the signals detected were at least predominantly due to a hybridization artifact. One interpretation of this artifact might be that it arose because of the intramolecular base-pairings in the globin mRNA. This would allow that even radioactive sense sequences might make interactions in a northern analysis with unlabeled sense RNA, even under conditions (hybridization and subsequent washing at 65°C) that we would consider stringent. However, a more likely explanation is that as an artifact of in vitro transcription, the radioactive RNA probes also contain trace amounts of the opposite strand; thus, when used in hybridization in the absence of appropriate specificity controls, one might detect species that would be incorrectly interpretted as AS RNA species.

AS, antisense; PHZ, phenylhydrazine; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

This paper was accepted for publication dependent on its association with the following commentary from one of the peer reviewers. The commentary can be accessed here:

Endoplasmic reticulum (ER) is a compartment simultaneously involved in the processes of protein synthesis and Ca²⁺ homeostasis in eukaryotic cells. Nascent polypeptide chains entering the secretory pathway, as well as extracytosolic portions of proteins destined for the external or internal cell membranes, translocate into the ER lumen. Within the ER, posttranslational processes including folding, glycosylation, subunit assembly and transfer to the Golgi compartment take place, assisted by chaperone proteins residing in the ER lumen [1,2]. In addition, proper protein folding and maturation depends on the maintenance of the oxidative milieu and high Ca²⁺ concentrations within the ER [3-5]. Reducing agents, depletion of ER lumenal Ca²⁺, inhibition of glycosylation or interference with the secretory pathway (by preventing transfer to the Golgi system), each lead to an accumulation of misfolded protein intermediates and increase the demand on the chaperoning capacity. Such conditions, collectively termed ER stress, have been shown to induce ER-specific stress response pathways. Two ER stress pathways concerned with protein processing are recognized today. The Unfolded Protein Response (UPR) was originally described in yeast [6], and has been more recently demonstrated in mammalian cells [7]. In mammals, the proximal ER stress-sensing element of this pathway appears to be Ire1p, a 110 kDa, highly conserved protein spanning the ER membrane. It is thought that when misfolded polypeptide chains accumulate within the ER causing a decrease in the level of free chaperones, the lumenal N-terminal portion of Ire1p undergoes dimerization. This in turn leads to a trans-autophosphorylation of the Ire1 cytosolic domains, triggering an ER-to-Nucleus (ERN) signaling [7]. The details of the ERN signaling are not understood at present. It appears to involve additional activities, in part residing in the cytosolic portion of the Ire1p molecule itself (RNA endonuclease activity), and in part contributed by proteins recruited through binding to Ire1p cytosolic domain (e.g. TRAF2, [8]). Overall, the major result of the UPR induction is a transcriptional upregulation of a number of stress proteins, including members of the family of Glucose Regulated Proteins (GRPs) such as GRP78 (BiP) and GRP94 (endoplasmin), as well as other chaperones (calreticulin, protein disulfide isomerase, ERp72) [7], thereby responding to the increased demand for the folding capacity within the ER.

Another ER-specific stress response has been termed Endoplasmic Reticulum Overload Response (EOR) [9]), and is triggered by certain of the same conditions known to activate UPR (e.g. glucose deprivation, glycosylation inhibition), as well as by heavy overexpression of proteins within the ER. The distinguishing feature of EOR is its association with the activation of the transcription factor NF-κB [9].

While necessary for the folding and maturation processes within the ER, the high concentration of Ca²⁺ ions in this compartment (1000-10,1000 times higher than in the cytosol [10,11]) is also a prerequisite for the universally employed Ca²⁺ signaling through ER Ca²⁺ channels (inositol 1,4,5-trisphosphate and ryanodine receptors) [12,13]. This high ER Ca²⁺ concentration is maintained by ER Ca²⁺ pumps, members of the family of Sarco-Endoplasmic Reticulum Calcium ATPases (SERCAs) [14-17]. Depletion of ER Ca²⁺ has been shown to cause a transcriptional induction of the mRNA for SERCA pumps [18]. However, the upregulation of an actual SERCA activity in such circumstances has not been demonstrated. In the present work, we provide the first functional evidence that depletion of ER Ca²⁺ during cell culture leads to an increase in the SERCA pumping capacity, as expressed by an enhancement of SERCA-mediated ⁴⁵Ca²⁺ uptake to microsomes isolated from Ca²⁺ -depleted PC12 cells. In addition, we show that similar functional SERCA upregulation may also be elicited by conditions of ER stress previously shown not to be accompanied by Ca²⁺ depletion. Taken together with our earlier results [19], the present data support the existence of a linkage between the induction of UPR and the regulation of SERCA gene expression.

This work shows for the first time an upregulation of ER Ca²⁺ -ATPase (SERCA) activity upon a disturbance of the internal ER environment by several agents known to evoke the ER stress response, or UPR. Since UPR may be activated by a variety of unrelated stimuli, e.g. Ca²⁺ depletion, inhibition of glycosylation, misfolding due to faulty amino acid precursors, energy deprivation or interference with the ER oxidative milieu [7,9], it may be thought of as a "final common pathway" for signaling ER stress caused by a range of factors. Therefore, we suggest that the UPR pathway is a likely candidate for the common mechanism to underlie the upregulation of SERCA activity following treatments with agents as different in their modes of action as EGTA, tunicamycin and DTT. Within this proposed common mechanism for SERCA activation due to UPR, some variation in the response amplitude between the individual agents appears to exist (Fig. 2). Further work will be required to understand the level (transcriptional and/or translational), and the exact mechanisms of such differences.

We have recently reported that treatment of PC12 cells with EGTA, tunicamycin, DTT or brefeldin A induced approximately 3- to 4-fold increases in the mRNA for SERCA2b, and approximately 2-fold increases in SERCA2b immunoreactive protein, with no corresponding changes for SERCA3 [19]. An upregulation of SERCA activity was also supported in this earlier work by an increase in the amount of SERCA phosphoenzyme. However, we would like to stress that despite these indications, the actual demonstration by the present paper of the stress-induced increase in the capacity of Ca²⁺transport into the ER was important for at least 2 reasons. First, as previously discussed [19, p.22371], technical difficulties precluded an accurate comparison of the extent of ATPase upregulation as measured at the immunoprotein and the phosphoenzyme level. Second, we were concerned with the possibility that ER stress, even if causing an increased synthesis of the immunoreactive protein, might at the same time by its very nature result in a faulty posttranslational processing of SERCA2b, with a degree of functional impairment. Since, as shown by the mutagenesis work [36], it is possible to disrupt the Ca²⁺ transport cycle of Ca²⁺ -ATPase without affecting at least some of the partial phosphorylation reactions, an actual full measurement in the present paper of the thapsigargin-sensitive Ca²⁺ transport into the ER was necessary to unequivocally demonstrate the validity of the ER stress-induced upregulation of SERCA2b activity.

PC12 cells were kindly provided by Drs. M. Grønborg and S. Gammeltoft, Glostrup County Hospital, Denmark. Dulbecco's modified Eagle's medium (DMEM;GlutaMAX-type), horse serum and fetal bovine serum were from GibcoBRL Life Technologies. ⁴⁵CaCl₂ (specific activity 0.2-2 Ci/mmole Ca²⁺) was from Pharmacia Amersham DK. Thapsigargin was a generous gift from Dr. S.B. Christensen, The Royal Danish School of Pharmacy.

Ca²⁺-ATPase, Ca²⁺-dependent adenosine triphosphatase; DMSO, dimethyl sulfoxide; DTT, dithiotreitol; EGTA, ethyleneglycolbis(β-aminomethyl ether)-N,N'-tetraacetic acid; EOR, endoplasmic reticulum overload response; ER, endoplasmic reticulum; ERN signaling, ER-to-nucleus signaling; GRP, glucose-regulated protein; SERCA, sarco-endoplasmic reticulum calcium ATPase; UPR, unfolded protein response

Two known enzymes catabolize the essential amino acid tryptophan in mammals. Tryptophan 2, 3 dioxygenase (TDO) is expressed predominantly in hepatic tissues and was the first inducible enzyme system discovered in mammals [1]. It controls serum tryptophan homeostasis and is induced following ingestion of tryptophan. A second enzyme, IDO, is distinguished from TDO by its expression pattern, substrate specificity and inducibility. IDO is expressed in a variety of non-hepatic tissues, including placenta, lung, gut and epididymis [2,3,4]. Except for the last named tissue where IDO is expressed constitutively, IDO is inducible by inflammatory mediators, including interferons. In addition, IDO catalyzes the breakdown of a variety of compounds which contain an indole ring, including D-tryptophan and serotonin, marking another difference from TDO, which is specific for L-tryptophan. Curiously, it appears as if tryptophan itself cannot induce IDO synthesis [5]. IDO is also suggested to be the evolutionary ancestor of certain novel myoglobins which occur in molluscs, marking IDO as an evolutionarily primitive enzyme [6].

IDO is known to be expressed in cells infected with intracellular pathogens such as Toxoplasma and Chlamydia species and also by viruses [7,8,9,10]. In the case of Toxoplasma and Chlamydia it has been proposed that IDO induction is a cellular defense mechanism, designed to limit the proliferation of the invading pathogen by depleting the essential amino acid tryptophan. IDO expressed in monocyte derived macrophages has also been found to inhibit the growth of extracellular bacteria such as group B streptococci [11], and is also induced in tumors taken from cancer patients [12]. In all of these systems the proximal inducer of IDO activity is interferon-γ (IFN-γ). Response elements for this cytokine have been identified in the human IDO promoter and have been shown to be essential for IFN-γ induction of reporter gene expression in vitro [13,14,15].

The unusual tissue distribution of IDO suggests that combating infection is not its only function. Our interest in IDO arose when we observed that tryptophan depletion was responsible for macrophage-induced inhibition of T cell proliferation in vitro [16]. Furthermore, we reported that a pharmacologic inhibitor of IDO, 1-methyl tryptophan, induced maternal rejection of allogeneic but not syngeneic murine fetuses [17]. As IDO is strongly expressed at the maternal-fetal interface in pregnant mice and women, we have suggested that IDO plays a role in fetal defense against the maternal immune system and could represent a novel means of immunoregulation. The apparently diverse functions and tissue distribution of IDO may have as a common theme the fact that tryptophan is the rarest essential amino acid and could be the target for cellular regulatory mechanisms. If so, tryptophan concentrations in cellular microenvironments might play a critical role in modulating various cellular processes in a way that cannot be achieved by the hepatic enzyme TDO which regulates systemic tryptophan concentrations.

The IDO promoter contains a diverse collection of motifs together with the IFN-γ response elements. These include motifs for transcription factors that bind to collagenase and elastase genes and motifs for the transcription factor MEP-1, which regulates transcription from the stromelysin-1 (MMP-3) and metallothionein genes [18,19]. Matrix metalloproteinases (MMPs) are responsible for modification of the extracellular matrix and are involved in wound healing, tumorigenesis, pregnancy and inflammation. In general, they regulate how cells interact with each other and with the extra-cellular matrix. Evidence for a tryptophan-reversible inhibition of MMP expression by IFN-γ has previously been presented [20, 21], although the exact mechanism is unclear. Therefore we decided to directly test whether IDO plays a role in controlling interactions with other cells and also the surrounding extracellular environment.

We have identified cells expressing IDO in vitro and used IDO antisense constructs to inhibit this expression. In addition, we have constitutively overexpressed IDO in adherent and non-adherent cell lines in vitro. Our results demonstrate that tryptophan catabolism has significant effects on cell adhesion and regulates the activity and expression of cyclooxygenases 1 and 2 (COX-1 and -2).

Tryptophan catabolism by cells expressing IDO is something of an enigma and has resulted in speculation as to why the body requires two enzymes with different tissue specificities to degrade the rarest essential amino acid [5, 34]. The inability of IDO to be induced by its own substrate exemplifies this puzzle. While IDO's role in controlling intracellular pathogens is well documented, there is little understanding of the reasons for IDO expression at sites in the body unlikely to be related to this function, such as the epididymis. The data we present here reveal that IDO expression is an important determinant of the way in which cells interact with their extracellular environment in vitro. In particular, cell adhesion is altered dramatically by overexpressing IDO in cells which do not otherwise express it, or inhibiting IDO expression in cells in which it is naturally induced following cell passage. Specifically, overexpression of IDO in RAW and MC57 cells resulted in the growth of macroscopic foci and other phenotypic alterations. The cell foci were multicellular aggregates, which grew vertically as well as horizontally across the plate surface and contained significant numbers of necrotic cells within their interior, as judged by trypan blue exclusion. Conversely, in P19 cell aggregates in which IDO expression was inhibited, there was a more dispersed phenotype with cells losing the ability to interact with each other. We have recently confirmed that IDO expression in RAW cells following cell passage is likewise important for correct cell adhesion (results not shown).

Our data support the hypothesis that IDO-induced alterations in PG synthesis can modify cell adhesion. We observed changes in the relative amounts of PGs in IDO-transfected RAW cells and reversal of the effects of IDO-expression by PG F_2α while PG E₂ stimulated focus formation. COX-2 was upregulated in IDO-expressing RAW cells. Similar effects of PG E₂ on cell morphology have recently been reported in the human embryonic kidney cell line HEK 293, which overexpressed COX-2 and PG E₂ synthase [35]. COX-2/PG E₂ synthase-expressing cells were highly aggregated, piled up and exhibited round shape morphology similar to the RAW cells described here. MC57 cells, which demonstrated similar changes in cell adhesion to RAW cells following IDO expression, exhibited lower levels of COX-2 synthesis upon IDO expression and lower levels of PG E₂ relative to other PGs such as D₂. Furthermore, adding back PGE₂ to IDO-expressing MC57 cells did not reproduce the wild type phenotype (not shown). Thus, similar effects on cell morphology were produced by opposite effects on COX-2 expression and PG E₂ production in these two cell lines. We are presently attempting to determine if products of COX-2 activity other than E₂ may be responsible for IDO effects in this cell line.

The mechanism of PG-induced changes in cell adhesion and morphology may involve MMP activity. Synthesis of MMPs such as collagenase I (MMP-1), gelatinase B (MMP-9) and matrilysin (MMP-7) has been shown to be dependent on the synthesis of PG E₂, suggesting that alterations in MMP expression may be instigated by alterations in PG synthesis [27,28,29,30]. Furthermore, both COX-1 and COX-2 have recently been shown to mediate adhesion of various cell types in vivo and in vitro [36, 37]. Thus, one possibility is that alterations in MMP expression and activity could modify cellular interactions with the extracellular matrix following IDO expression. Consistent with this possibility is the observation that MMP expression in P19 cells was correlated with the degree of IDO-antisense expression. RAW transfectants over-expressing IDO showed reduced expression of collagenase I (MMP-1), and also bound less well to collagen-coated plates than controls. Collagen is one of the principal components of the extracellular matrix and RAW cells bind poorly to fibrillar type I collagen unless it is denatured or activated by collagenase [38]. Thus, the diminished expression of MMP-1 in IDO transfectants could explain their weaker binding to this substrate. The mechanism by which IDO regulates prostaglandin synthesis is yet to be determined. Tryptophan is a stimulatory co-factor for COX and degradation of tryptophan in the intracellular environment could alter COX activity. Alternatively, IDO might influence COX activity and expression through competition for or release of heme, which both enzymes require. In vitro, arachidonic acid stimulates the dissociation of heme from IDO and this correlates with IDO stimulatory effects on COX [39], providing circumstantial support for the latter possibility.

The alterations in COX-2 expression observed in IDO-expressing RAW and MC57 cells are a particularly interesting feature of our results. COX-2 is inducible by a number of inflammatory mediators including IFN-γ [40] and LPS [32]. These also induce IDO. Treatment of IDO-expressing RAW clones with a known inducer of COX-2 (LPS), revealed a lack of correlation between COX-2 RNA and protein levels. Clones 11 and 22 showed low levels of COX-2 message but high levels of protein following LPS treatment. This suggests that COX-2 RNA and/or protein turnover may be affected by IDO expression. Other workers have noted that non-steroidal anti-inflammatory drugs, which inhibit COX activity result in increased COX protein expression [41], while differences between COX protein expression and activity have been reported to be produced by some cytokines, including tumor necrosis factor-α, and also nitric oxide donors [42, 43]. Although not well understood, evidence for regulation of COX expression at the post-transcriptional level is increasing [44, 45]. The down regulation of COX-2 transcripts in LPS-treated, IDO-expressing RAW cells is reminiscent of the endotoxin tolerance effect observed in human THP-1 promonocytic cells. Cells pretreated with LPS and thus expressing COX-2 showed down regulation of COX-2 mRNA when subjected to a second LPS exposure [46]. In addition, the COX-2 inhibitor flufenamic acid induced COX-2 expression in RAW cells but inhibited TNF-α or LPS-induced COX-2 expression in the same cell type [47].

As both MMPs and COX-2 are important factors in tumor development [48, 49] IDO's role in tumorigenesis bears investigating. We have observed IDO expression routinely in murine tumors in vivo, and are presently investigating the growth properties of tumors with altered IDO expression. In addition, our recent work indicates a role for IDO during pregnancy. Pharmacological inhibition of IDO results in pronounced inflammation, complement activation and fetal loss [17, 50]. Prostaglandins may provide a common link between these important biological phenomena.

IDO regulates adhesion of cells to normal growth substrates. In so doing it modulates the expression and activity of COX-2 and certain MMPs. RAW cells and MC57 cells overexpressing IDO grew as multicellular foci. In the case of RAW cells, this was due to elevated PGE relative to other prostaglandins. P19 cells in which endogenousIDO expression was disrupted by antisense expression, showed lower adhesiveness. Thus, tryptophan catabolism exerts control over fundamental cellular functions.

CMV: cytomegalovirus

COX: cyclooxygenase

IDO: indoleamine 2,3 dioxygenase

IFN-γ interferon gamma

LPS: lipopolysaccharide

MMP: matrix metalloproteinase

NSAID: non-steroidal anti-inflammatory drug

PG: prostaglandin

TDO: tryptophan 2, 3 dioxygenase