Supercoils in plant DNA: Nucleoid sedimentation studies

It is well known that the final structural organization of eukaryotic DNA in the nucleus is mediated by several levels of folding of the long linear DNA molecule. The primary folding responsible for a six- to sevenfold compaction of DNA is the winding of the double helix around the nucleosome core formed by an octamer of the nucleosomal histones (two molecules each of H2A, H2B, H3 and H4) (Klug et al. 1980). At a second level the nucleosome fibre (10 nm) is further folded into thicker fibres (30 nm) organized as supercoils or super-beads, stabilized by histone Hl (Finch & Klug, 1976; Renz et al. 1977). These two levels of folding are still not sufficient to bring about the final compactness of the DNA in the interphase nucleus and more particularly in the metaphase chromosome. It is believed that the thicker fibres are organized in topologically constrained DNA loops anchored to insoluble nuclear matrix structures. The matrix is isolated by high salt extraction of nuclei, followed by nuclease treatment (Berezney & Coffey, 1974, 1977) and consists of a characteristic set of proteins, whose complexity depends on the isolation conditions (for review see Kaufmann et al. 1986). The constrained loop concept comes from microscopic studies and from investigations on histone-depleted nuclei.

Nuclear structures devoid of the nuclear membrane and of histones by non-ionic detergent and high salt (2M-NaCl) treatment, called nucleoids (Cook & Brazell, 1975, 1976; Cook et al. 1976), nucleosome-free interphase chromosomes (Hancock & Hughes, 1982) and high salt resistant nuclear structures (Mullenders et al. 1982), have been studied in different animal systems. These structures preserve the typical nuclear morphology and behave in a manner characteristic of intact circular supercoiled DNA. This refers to the way the sedimentation behaviour of the nucleoids changes as a function of the ethidium bromide (EthBr) concentration in sucrose density gradients and to the alterations in the sedimentation rate observed following introduction of single-strand breaks in the DNA by gamma irradiation (Cook & Brazell, 1975, 1976; Cook et al. 1976).

When nucleoids are sedimented in a series of gradients containing increasing concentrations of the intercalating dye EthBr, their sedimentation rates are altered biphasically, the rate first falls and then increases again. This behaviour is explained (by analogy with the effect of EthBr on covalent close circular DNA; Vinograd et al. 1965) by the structural changes due to EthBr intercalation, which first reduces the number of negative supercoiled turns, then results in a structure with no net supercoiling and further creates positive supercoiling (for review see Mattern, 1984). Irradiated DNA, on the other hand, behaves as if it no longer contains superhelical DNA, i.e. it behaves like nicked DNA (Cook & Brazell, 1975). Both these features of the nucleoid, its sedimentation behaviour upon EthBr titration and following gamma irradiation, have been interpreted as an indication that eukaryotic DNA is supercoiled. Interestingly, similar observations have been reported on the so-called folded Drosophila genomes prepared in 0·9M-NaCI, i.e. on nuclear structures lacking only histone Hl but preserving the rest of the histones (Benyajati & Worcel, 1976). These observations have also been interpreted as showing the presence of supercoils in Drosophila DNA loops.

No similar studies aimed at the elucidation of the higher order organization of nuclear DNA in plants have been conducted. Plants share many common features with the other eukaryotes, but there are also many peculiarities distinguishing them. For example, they possess unusually large genomes, highly variable in size and organization from species to species, a lot of repetitive DNA, low relative amounts of transcribed sequences, variable ploidy levels (even within a single plant or tissue), flexibility and plasticity encountered nowhere else (for reviews see Flavell, 1980, 1982; Sorenson, 1984; Walbot & Cullis, 1985). The nucleo-somal level of organization of the genetic material seems to be the same as in other eukaryotes (for reviews see Nagi, 1982a; Spiker, 1984, 1985). Nothing, however, is known about the secondary and tertiary levels of DNA folding and also about the relationship between DNA structure and function. In view of the complete lack of knowledge concerning these issues we have undertaken a series of investigations, the first of which is reported here. Using the nucleoid sedimentation method devised by Cook & Brazell (1975, 1976) we show that plant DNA is also supercoiled. In addition, we show that nucleoid sedimentation in plants gives rise to two distinct nucleoid bands; their nature remains at present unclear and is the subject of further studies.

Because plant cells possess thick walls the direct use of cells in the nucleoid sedimentation assay was not possible and isolated nuclei were used instead. Nuclei were obtained according to the procedure described by Ivanchenko et al. (1987) either from whole maize (line M32O) embryos germinated for different periods of time or from seedling root meristems. Germination was performed on moist filter paper at 28°C in the dark and the embryos were dissected by hand. Microscopic observation of the nuclear pellet showed the presence of intact nuclei contaminated with some cell debris and small amounts of starch grains. In some cases nuclei from Hordeum vulgare and Glycine max were isolated by the same procedure.

All manipulations were carried out at 4°C. Nucleoids were obtained according to Cook & Brazell (1975, 1976) with some modifications (Mullenders et al. 1983). Linear sucrose gradients (30 ml, 15% −30% sucrose, containing 2 M-NaCl, 1 mM EDTA, 25mM-Tris HCl, pH 8·0 and variable amounts of EthBr) wre prepared on 1 ml 65% sucrose cushions containing 0·4gml⁻¹ CsCl and were overlaid with l·5ml of lysis solution (05% Triton X-100, 2M-NaCl, lmM-EDTA, 25 mM-Tris HC1, pH8·0). About l×10⁶ isolated nuclei resuspended in 0’5 ml of water were lysed on the top of the gradients in the dark at 4°C for 30–60 min. The centrifugation was performed in a Sorvall TV-850 ultra vertical rotor for 4h at 35 000rev. min⁻¹ at 4°C. In some cases a Sorvall AH-650 swinging bucket rotor was used. The position of the nucleoid bands was determined under 254 nm u.v. light. The sedimentation rate of nucleoids was expressed relatively to the sedimentation behaviour of respective controls as specified in the text. Alternatively, gradients were analysed using a Uvicord fraction collector and recorder operated at 254 nm (LKB, Sweden). The position of the bands was measured as the distance from the centre of the absorbance peak to the bottom of the tube and expressed as a percentage of the distance covered by the control nucleoids. Both methods gave comparable results.

To analyse the DNA and protein components of the nucleoids, they were collected automatically as described above; the respective bands from several gradients were pooled and dialysed overnight in the dark at 4°C against 30—50 volumes of TE buffer (10mM-Tris HCl, pH8·0, 1 mM-EDTA, 2 changes). During the dialysis step against this low ionic strength buffer the proteins dissociated from the DNA and formed insoluble aggregates, which could be recovered by low speed centrifugation. After removing the proteins the supernatant containing practically all nucleoid DNA was precipitated with cold ethanol and treated with RNase A (40 μgml⁻¹) and proteinase K (100μgml⁻¹) for 1 h at 37°C. The samples were then subjected to a standard phenol extraction procedure. Finally the samples were dissolved in small volumes of TE buffer and stored at −20°C for further analysis. Total nuclear DNA was obtained from isolated nuclei by phenol extraction preceded by lysis of the nuclei in the buffer specified above.

Native and denaturing agarose gel electrophoresis (0·8% agarose) of the purified total nuclear or nucleoid DNA was performed according to Maniatis et al. (1982): γ DNA or its Pst1 digest were used as molecular mass markers. Restriction enzyme digestions were performed as described by Maniatis el al. (1982) or according to the instructions of the manufacturer (Promega Biotech, USA).

Proteins were subjected to polyacrylamide gel electrophoresis in slabs containing sodium dodecyl sulphate (Laemmli, 1970).

G + C content was estimated based on the equation of De Ley (1967) following determination of E₂₆₀/E₂₈₀ as specified by Fredericq el al. (1961).

A gamma source (GUBE 4000, USSR) containing ⁶⁰Co was used to deliver a 10 rad min⁻¹ dose rate over a uniform field to final doses of up to 1000 rad as specified in the text. For the analysis of the gamma-irradiated material dry seed embryos or embryos germinated for 16 h were dissected from the seeds, placed in the homogenization buffer and irradiated at 4°C before isolation of the nuclei. Alternatively, irradiation was performed on isolated nuclei suspended in homogenization buffer.

The flow cytometric analysis was performed on isolated nuclear suspensions following the procedure described by Christov & Yantchev (1985). Nuclei were collected by low speed centrifugation, transferred into staining solution containing 12·5μgml⁻¹ EthBr, 25μgml⁻¹ mitramycin and 7·5mM-MgCl₂ and incubated for 15 min at 24°C. The staining was followed by a brief RNase treatment at room temperature. Between 10³ and 10⁵ nuclei were analysed using an ICP-11 impulse cytophotometer (Phywe, Göttingen, West Germany).

Nucleoid samples were layered on 0·l M-sucrose cushions containing 2M-NaCl and 1% paraformaldehyde and centrifuged in a microcentrifugation chamber for 5 min at 3500g on freshly glow-discharged carbon-coated grids. The specimens were stained in 1% phosphotungstic acid in 70% ethanol and rinsed in absolute ethanol (Miller & Bakken, 1972). Observations were made with a JEM-100B electron microscope at 80kV.

During the initial phase of our experiments we optimized some of the conditions of the nucleoid sedimentation assay. Thus, best results were obtained with the lysis solution specified in Materials and methods and lysis times between 30 and 60 min. It was also essential to use freshly isolated nuclei as storage of nuclei led to some smearing of the bands.

Fig. 1 shows a centrifugation tube following sedimentation of nucleoids formed from nuclei of embryos germinated for 16 h. Two distinct nucleoid bands were observed under u.v. illumination; the presence of the two bands was confirmed upon automated fractionation of the gradients (Fig. 1). The same phenomenon was observed with nuclei from all other maize tissues tested (dry embryos, root tip meristems, leaves) and more importantly with nuclei isolated from Hordeum vulgare root meristematic cells and Glycine max dry embryos (data not shown). It has to be pointed out that the relative amounts of the bands varied with the time of germination. In the dry embryo material the two bands were present in almost equal amounts. More or less the same situation was observed at 16 h of germination, but when nucleoids were obtained from root meristems the intensity of the upper band was much less than that of the lower one.

As a control purified DNA, either total DNA or from the individual nucleoid bands, was run in parallel tubes. All three DNA samples sedimented as single peaks at nearly the same positions, the position of the upper nucleoid band. This could mean that the upper band represented some pure DNA released in the course of the isolation and centrifugation procedures. That this was not the case was evident from all further data on the protein components of the two bands, the electron microscope pictures and EthBr titrations (see below). Thus, the upper material is a bona fide nucleoid band characteristically present in all plant sources tested.

It is known that restriction enzymes with recognition sequences CCCGGG or CCGG (Smai, HpaII, PstI) do not cleave if the C residue to the 3′ side of the cutting site is methylated (Mann & Smith, 1977). As 20–25% of the cytosine residues in most higher plant genomes are methylated (Bedbrook et al. 1978), these enzymes are not expected to cleave the nuclear DNA effectively. However, they readily cleave plant extra-nuclear DNA, giving rise to characteristic ladders of fragments (for an example of a PstI digest of chloroplast DNA see Teemuteeri & Lokki, 1984). This difference in the response of the nuclear and extranuclear genomes to digestion with the above-mentioned enzymes was used to check whether the two bands were both of nuclear origin or whether one could be due to the large amounts of cytoplasmic DNA present in the plant cell (see e.g. Flavell, 1982). Indeed, as can be seen from Fig. 2 there was almost complete failure of PstI to cleave DN As from both the lower and the upper band and total nuclear DNA. As a control for the reaction, digestion of intact γ DNA gave the known ladder of restriction fragments (Fig. 2, track 2). The low degree of cleavage observed for the two bands shows that both contain mainly nuclear DNA.

The presence of two nucleoid bands could be due to some variations in the size, shape and/or the density of the nucleoids. Electron microscopic examination of the two bands revealed a typical appearance, an electrondense inner mass out of which numerous long DNA fibres extended to form a dense network interwoven to different degrees (Fig. 3). The nucleoids of the two bands did show some differences in compactness and size, the upper band nucleoids being somewhat smaller; statistically valid conclusions can however be drawn on the basis of more experiments. Some differences could also be seen in the appearance of the DNA halo (compare Fig. 3A and B).

The possibility exists that the two bands reflect heterogeneity of the starting nuclear populations with respect to ploidy levels. That the observed two nucleoid bands were not due to heterogeneity in the amount of DNA per cell (over that due to different cell cycle phases) is clear from the observation that two bands were obtained from root meristematic cells. These cells formed a homogeneous dividing population with 2C G₁ content as evidenced from their distribution histogram according to the DNA content (Fig. 4). It is known, however, that even in such homogeneous populations there are several nuclear phenotypes differing in size and in the amount and localization of the heterochromatic regions (Vidal et al. 1984). The possibility that the two bands reflect some of these differences remains to be elucidated. Our preliminary attempts to fractionate the total nuclear suspension into different subtypes by the use of stepwise sucrose gradients have been unsuccessful.

The dry maize embryo is known to contain both 2C and 4C nuclei; moreover, in any population of seeds there could be variations among seeds depending on how the metabolism of any seed was arrested during development and drying (Bewley & Black, 1978). Our flow cytometric analysis of dry embryo nuclear populations confirmed literature data in that both 2C and 4C nuclei were present. The proportion of 4C nuclei was much higher than that in the meristem (not shown). Again, no nuclei of higher than 4C DNA content were present. This observation confirms the conclusion drawn from the experiments with the meristematic cells that the two nucleoid bands are not due to ploidy heterogeneity.

As can be seen from Fig. 5 the intercalating agent EthBr has a significant effect on the sedimentation properties of the nucleoids. Only the behaviour of the lower (major) band is presented; the upper band behaves in a similar manner. As the concentration of the intercalating dye increased, the distance travelled by the nucleoid fell to a minimum and then increased again, the sedimentation rate approaching that of the control nucleoids, i.e. there was a characteristic biphasic alteration in the sedimentation rates (Fig. 5, upper curve). There is however a substantial feature distinguishing the plant nucleoid behaviour from that observed in the animal systems and it concerns the magnitude of the difference between the maximum sedimentation rate (at very low EthBr concentrations) and its minimum (around l–2·5μgml⁻¹ EthBr). For the animal nucleoids the minimal sedimentation rate is about 0·5 of the maximum (Cook & Brazell, 1975) whereas in our experiments with the 15% −30% sucrose gradients it never decreased below 0·85–0·87 of the maximum. In less dense gradients of the same steepness (10% −25%) the minimum sedimentation rate was 0’82 of the maximum; the use of less steep gradients (15%—25%) did not lead to a significant change in the value. The sedimentation rate of the irradiated material was also relatively high (see below).

Gamma rays are known to induce single-strand scissions in the DNA and irradiation of superhelical DNA leads to abolition of the biphasic response of the sedimentation behaviour to increasing EthBr concentrations, i.e. irradiated superhelical DNA behaves as a relaxed molecule. Indeed, when the embryos were irradiated the biphasic response was completely lost (Fig. 5, lower curve) (in this case unirradiated material was used as a control).

It should be noted that the doses that produce these effects were very small; a similar effect to that presented in Fig. 5 (100 rad) was obtained even when 50 rad were used. Moreover, the dose-response curve reflecting the changes in the sedimentation behavior as a function of the dose of irradiation flattened off at 200 rad (Fig. 6) while similar curves with mouse thymus cells and human lymphocytes continuously dropped to at least 1000 rad (Weniger, 1982; L. M. Stoilov, unpublished results).

The existence of two nucleoid bands in the plant material was totally inexplicable, bearing in mind the accepted notion that the nucleoids did not sediment as separate particles but as one aggregate (Weniger, 1982; Mattern, 1984). Such a notion was based on the observation that even a mixture of irradiated and nonirradiated cells gave only one DNA peak (Weniger, 1982). In order to check whether the aggregate concept was also applicable to plants we performed the following mixing experiment. Control nuclei and nuclei exposed to 500 rad gamma irradiation were mixed in a 1:1 ratio and applied to the top of a gradient; separate tubes contained only control and only irradiated material. As predicted from the aggregate concept the mixed nuclei populations gave only two (not four) nucleoid bands with intermediate sedimentation rates and of the usual widths (Table 1). Thus it seems that the aggregate concept is also valid for plants but mixed aggregates are only formed between otherwise identical bands whose DNA differs only in the content of single-strand breaks. Evidently the upper and the lower bands represent totally distinct nucleoid entities, so different from one another that they always sediment as individual bands.

When proteinase K and RNase A treatment was performed during the lysis of the nuclei on top of the gradients the sedimentation behaviour of the nucleoid structures changed significantly. When proteinase K was present in the lysis solution the nucleoid bands formed became more diffuse and their sedimentation rate decreased to the values characteristic of the irradiated material. Even more drastic was the effect of RNase. This treatment done at relatively high enzyme concentration led to the complete loss of the nucleoid bands in the tubes (data not shown). Hence this treatment totally destroyed the integrity of the nucleoids.

DNA from the two bands was isolated by phenol extraction and compared by agarose gel electrophoresis. As can be seen from Fig. 2 there were no distinct differences in the electrophoretic properties of the DNA derived from the two nucleoid bands and of the total nuclear DNA, run on native gels. All DNAs were found in the region of 50 kb. Agarose gel electrophoresis under denaturing conditions also failed to show any significant differences in electrophoretic behaviour (not shown). As already mentioned the purified DNAs from the two bands sedimented with similar rates through sucrose gradients.

Attempts to find differences in the G + C content of the DNA of the two bands were also unsuccessful. The ratios of absorbances E₂₆₀/E₂₈₀ were determined and the G + C content was calculated according to the equation of De Ley (1967). The values obtained were 45·8% for the upper band DNA, 5l·6% for the lower band DNA and 46·5% for the total nuclear DNA. Bearing in mind the low accuracy of the method we can infer that the G + C contents of the two bands do not show significant differences.

The proteins present in the nucleoid bands were obtained by dialysis against the low ionic strength TE buffer (see Materials and methods). Under these conditions the majority of the nucleoid proteins of molecular mass 45–67K (K = M_r× 10⁻³) dissociated from the DNA as judged by the fact that the DNA in the buffer contained only trace amounts of these proteins. The dissociated protein was analysed in 12·5% polyacrylamide gels containing sodium dodecyl suphate (Fig. 7). Molecular mass markers and total acid-soluble chromatin proteins were run for comparison. As seen in Fig. 7 no proteins with the mobility of histones were present. Interestingly, the two bands showed exactly the same protein profiles characterized by the presence of only a few bands in the region of 45–67K. Proteins of similar molecular masses have been described in HeLa nucleoids (Cook et al. 1976; Adolph, 1980; Mullenders et al. 1982).

To study the higher order folding of DNA in the plant nuclei we have investigated the sedimentation behaviour of plant nucleoids in linear sucrose gradients containing EthBr. When titrated with increasing concentrations of EthBr the nucleoids show the biphasic changes in sedimentation rate characteristic of all circular supercoiled DNA. In addition, the biphasic response is completely lost upon induction of single-strand breaks in the DNA by gamma irradiation. These observations should be interpreted as showing that plant DNA is also organized in constrained loops that contain negative supercoils.

Another important feature concerns the EthBr concentration at the so-called equivalence point (the point at which EthBr intercalation removes all the negative superhelical turns and native closed circular DNA behaves as nicked DNA). The equivalence point in our experiments is in the range of 1–2 ·5μgml⁻¹, i.e. the same as that observed for SV4O (Crawford & Waring, 1967; Bauer & Vinograd, 1968; Mayer & Levine, 1972), γ (Hinton & Bode, 1975), Escherichia coli (Worcel & Burgi, 1972), D. melanogaster (Benyajati & Worcel, 1976) and different mammalian and chicken sources (Cook & Brazell, 1975, 1976; Cooket al. 1976). This shows that the superhelical densities in all these DNA molecules are the same, approximately one negative superhelical turn per 200bp (Benyajati & Worcel, 1976).

The significant differences between animal and plant nucleoids concern two points: the relatively high sedimentation rate at the equivalence point and of the irradiated material, and the low radiation doses at which the dose-response curve levels off. These features might reflect one and the same characteristic of the plant material. It is not unreasonable to assume, as Mullenders et al. (1983) have already done, that only very large DNA loops determine the sedimentation behaviour of the nucleoids and that the relative amount of these large DNA loops in the plant genome is low. An alternative plausible explanation might be a low relative content of superhelical DNA loops in the plant sources. Such a low relative content of superhelical loops might bear some relation to the relatively low proportion of transcribed and translated DNA sequences in plants; 0·1–1% of the total DNA in plants (Nagi, 19826; Flavell, 1982) versus 5–10% in animals (Pederson, 1978; Lewin, 1975; Mathis et al. 1980). It should be noted that the same results would have been obtained if the isolation procedure nicked most of the nuclear DNA. Although difficult to test, this possibility cannot be ruled out at present.

From studies on animal cell nucleoids it is known that the integrity of the structure depends on the presence of protein and RNA (for example see Cook et al. 1976). The same is evidently true for plant nucleoids as indicated by the proteinase K and RNase A treatments.

The sedimentation of high salt resistant nuclear structures of plants through sucrose gradients revealed unexpectedly the presence of two bands. Control mixing experiments (a 1:1 mixture of unirradiated and irradiated material run in a single tube) showed the presence of two bands of intermediate sedimentation rates, which confirmed the aggregate notion of the sedimentation behaviour of nucleoids (Weniger, 1982). In addition, they showed, however, that the two bands normally observed were so different from one another that it was not possible to form a single band from them.

Control experiments were run to check whether both nucleoid bands were of nuclear origin. The high methylation level of the C residues in the nuclear DNA was supposed to preclude its effective cleavage with particular restriction enzymes while chloroplast and mitochondrial DNA could be cleaved producing specific restriction fragment patterns. These experiments confirmed the nuclear origin of both nucleoid bands.

The morphology of the nucleoids was studied by electron microscopy. The plant nucleoids possess the typical appearance of their animal counterparts (e.g. see fig. 2 of Jackson et al. 1984). The structures in the two nucleoid bands differed mainly in size, the lower band containing larger structures.

Attempts were made to find some differences either in the DNA or in the protein content of the two bands. Isolated DNA was compared electrophoretically, on the basis of its G + C content and its sedimentation in sucrose gradients. In no respect could differences be detected. The same applied to the protein content of the two bands; the electrophoretic patterns were indistinguishable, characterized by the presence of only a few bands in the region of 45–67K. The proteins of animal cell nucleoids gave very similar electrophoretic patterns (Cook et al. 1976; Adolph, 1980; Mullenders et al. 1982). Some of the proteins have been characterized as lamins, the major components of the nuclear pore complex-lamina. Recently, similar proteins have been described in isolated plant nuclear matrices (Ghosh & Dey, 1986). On the basis of the similarities between the plant and animal nucleoid proteins it might be suggested that the proteins taking part in DNA loop formation are highly conserved. Our preliminary data using peptide mapping and immunochemical techniques do confirm a great degree of conservation of these proteins.

The existence of the two nucleoid bands in plants could reflect some heterogeneity of the nuclear populations with respect to size, shape and/or density. That the two bands are not due to different ploidy levels is evident from the observations that a dividing diploid population, as represented by the root meristematic cells, also gives two bands.

The possibility that cell cycle-dependent differences in nuclear parameters determine the presence of two bands cannot be excluded at present. HeLa cell nucleoids derived from synchronous G₁, S and M populations did show significant differences in the relative sedimentation rate, small S phase nucleoids sedimenting nine times faster than the larger mitotic nucleoids (Warren & Cook, 1978). While this difference is probably too great to allow separation of S and M nucleoids within a single standard gradient, such a separation would be expected for S and G₁ nucleoids whose sedimentation rate differed by only 20–30%. Nevertheless, nucleoids from unsynchronized populations never exhibited two bands (Cook & Brazell, 1975, 1976; Cook et al. 1976; Warren & Cook, 1978). The reason for this discrepancy remains unclear; the aggregate notion of nucleoid sedimentation behaviour could be a plausible explanation. If the two plant bands do reflect cell cycle-dependent differences, these should be highly specific for plants, as they do not allow the formation of a single aggregate out of the distinct phase nucleoids.

In summary, the nucleoid sedimentation studies performed in this work lead to the following main conclusions.

1. Plant DNA contains supercoiled loops with the same superhelical density as observed for phage, viral, bacterial, insect and vertebrate genomes. The relative amount of superhelical loops in the plant genome seems to be much lower than in the animal genome, correlating with its much lower transcriptional activity. The proteins responsible for the constraining of the DNA loops are very similar between plants and animals.

2. A characteristic feature of plant nuclear populations is that they form two distinct nucleoid bands, indistinguishable by all criteria used in this study. Further investigations at the cellular and molecular levels are required to clarify this point.

The authors thank Dr K. Christov for performing the flow cytometric analysis and Dr L. Todorova for supplying the dry maize seeds. This project has been completed with the financial support of the Committee for Science at the Council of Ministers under contract 486.