Changes in chromatin structure at the replication fork. II The DNPs containing nascent DNA and a transient chromatin modification detected by DNAase I.

Discrete deoxyribonucleoproteins (DNPs) containing nascent and/or bulk DNA, were obtained by fractionating micrococcal nuclease digests of nuclei form 3H-thymidine pulse (15-20 sec) and 14C-thymidine long (16 h) labeled sea urchin embryos in polyacrylamide gels. One of these DNPs was shown to contain the micrococcal nuclease resistant 300 bp "large nascent DNA" described in Cell 14, 259-267, 1978. The bulk and nascent mononucleosome fractions provided evidence for the preferential digestion by micrococcal nuclease of nascent over bulk linker regions to yield mononucleosome cores with nascent DNA. DNAase I was used to probe whether any nascent DNA is in nucleosomes. Nascent as well as bulk single-stranded DNA fragments occurred in multiples of 10.4 bases with higher than random frequencies of certain fragment sizes (for instance 83 bases) as expected from a nucleosome structure. However, a striking background of nascent DNA between nascent DNA peaks was observed. This was eliminated by a pulse-chase treatment or by digestion of pulse-labeled nuclei with micrococcal nuclease together with DNAase I. One of several possible interpretations of these results suggests that a transient change in nucleosome structure may have created additional sites for the nicking of nascent DNA by DNAase I; the micrococcal nuclease sensitivity of the interpeak radioactivity suggest its origin from the linker region. Endogenous nuclease of sea urchin embryos cleaves chromatin DNA in a manner similar to that of DNAase I.     

Reduced repeat length of nascent nucleosomal DNA is generated by replicating chromatin in vivo

Micrococcal nuclease digestion of nuclei from sea urchin embryos revealed transient changes in chromatin structure which resulted in a reduction in the repeat length of nascent chromatin DNA as compared with bulk DNA. This was considered to be entirely the consequence of in vivo events at the replication fork (Cell 14, 259, 1978). However, a micrococcal nuclease-generated sliding of nucleosome cores relative to nascent DNA, which might account for the smaller DNA fragments, was not excluded. In vivo [3H]thymidine pulse-labeled nuclei were fixed with a formaldehyde prior to micrococcal nuclease digestion. This linked chromatin proteins to DNA and thus prevented any in vitro sliding of histone cores. All the nascent DNAs exhibiting shorter repeat lengths after micrococcal nuclease digestion, were resolved at identical mobilities in polyacrylamide gels of DNA from fixed and unfixed nuclei. We conclude that these differences in repeat lengths between nascent and bulk DNA was generated in vivo by changes in chromatin structure during replication, rather than by micrococcal nuclease-induced sliding of histone cores in vitro.     

Maturation of newly replicated chromatin of simian virus 40 and its host cell

The DNA in replicating simian virus 40 chromatin and cellular chromatin was labeled with short pulses of [³H]thymidine. The structure of pulse-labeled nucleoprotein complexes was studied by micrococcal nuclease digestion. It was found that in both newly replicated viral and cellular chromatin, a structural state appears which is characterized by an increased sensitivity to nuclease and a faster than usual rate of cleavage to DNA fragments of monomeric nucleosome size and smaller. Pulse-chase experiments show that each of these effects requires a characteristic time to disappear in both systems, suggesting the existence of different sub-processes of chromatin maturation. One of these processes, detectable by the reversion of the unusually fast production of subnucleosomal fragments, is delayed in SV40 chromatin replication.     

Studies on the mode of segregation of histone nu bodies during replication in HeLa cells

Two models were tested for the mode of distribution of histone nu bodies at the replication fork. The replication fork was labeled by brief incubation of cells with ³H-thymidine. Nuclei were isolated and digested with low levels of micrococcal nuclease, and the kinetics of cleavage of the pulse-labeled chromatin DNA were compared to the kinetics of cleavage of parental chromatin DNA. In chromatin labeled for 30 sec to 10 min, the rate of cleavage of the pulse-labeled region into monomeric nu body-sized units exceeded the rate of cleavage of parental chromatin by a factor of 2, but did not approach the predicted value of 5–6 for random segregation. This value dropped to 1.6 in 15 min and was euivalent to parental chromatin in 20 min labeling experiments. DNA synthesized in the presence of cycloheximide was also digested at twice the rate of parental chromatin DNA.A Poisson analysis of the kinetics of cleavage by micrococcal nuclease further confirmed these observations. The predicted difference in the rate of production of monomeric, dimeric, and trimeric deoxyribonucleoprotein units was very similar to the experimental values of both total chromatin and nascent chromatin. Thus the nu body spacings in newly replicated chromatin closely approximate those in parental chromatin.These results agree well with a conservative or nondispersive model of nucleosome distribution in which the proteins are associated with one of the two daughter chromosomes during replication.     

Changes in chromatin structure at the replication fork. DNase I and trypsin-micrococcal nuclease effects on approximately 300- and 150-base pair nascent DNAs

DNase I, trypsin, and micrococcal nuclease are used to further probe the structure of nascent deoxyribonucleoprotein (DNP) fractions which appear after in vivo 20-s pulse labeling of sea urchin embryos with [3H]thymidine. We present evidence that the large nascent DNP which protects the approximately 300-base pair large nascent DNA consists of at least one nucleosome core. This is based on fractionation in denaturing polyacrylamide gels of DNA extracted from large nascent DNP fractions of a micrococcal nuclease + DNase I digest of nuclei. The data also suggest the existence of a DNase I-hypersensitive site(s) within the large nascent DNP; this is consistent with the hypothesis that the latter consists of closely packed dinucleosome cores. Histone H1 and non-histone proteins do not account for the previously reported unusual hyperresistance of the large nascent DNA against micrococcal nuclease. The protection offered this approximately 300-base pair nascent DNA was not eliminated by an 0.2-microgram/ml trypsin pretreatment which removes the above proteins from the chromatin. However, 5-10 micrograms/ml of trypsin, which remove a portion of the NH2 termini of the four core histones of nucleosomes, eliminate the hyperresistance of the large nascent DNA to subsequent micrococcal nuclease digestion, while nascent and bulk monomer DNAs remain unaffected. This indicates histone-histone and/or histone-DNA interactions within the large nascent DNP which differ from those of nascent and bulk mononucleosome cores.     

Structure of eukaryotic chromatin. Evaluation of periodicity using endogenous and exogenous nucleases.

DNA isolated from (a) liver chromatin digested in situ with endogenous Ca2+, Mg2+-dependent endonuclease, (b) prostate chromatin digested in situ with micrococcal nuclease or pancreatic DNAase I, and (c) isolated liver chromatin digested with micrococcal nuclease or pancreatic DNAase I has been analyzed electrophoretically on polyacrylamide gels. The electrophoretic patterns of DNA prepared from chromatin digested in situ with either endogenous endonuclease (liver nuclei) or micrococcal nuclease (prostate nuclei) are virtually identical. Each pattern consists of a series of discrete bands representing multiples of the smallest fragment of DNA 200 +/- 20 base pairs in length. The smallest DNA fragment (monomer) accumulates during prolonged digestion of chromatin in situ until it accounts for nearly all of the DNA on the gel; approx. 20% of the DNA of chromatin is rendered acid soluble during this period. Digestion of liver chromatin in situ in the presence of micrococcal nuclease results initially in the reduction of the size of the monomer from 200 to 170 base pairs of DNA and subsequently results in its conversion to as many as eight smaller fragments. The electrophoretic pattern obtained with DNA prepared from micrococcal nuclease digests of isolated liver chromatin is similar, but not identical, to that obtained with liver chromatin in situ. These preparations are more heterogeneous and contain DNA fragments smaller than 200 base pairs in length. These results suggest that not all of the chromatin isolated from liver nuclei retains its native structure. In contrast to endogenous endonuclease and micrococcal nuclease digests of chromatin, pancreatic DNAase I digests of isolated chromatin and of chromatin in situ consist of an extremely heterogeneous population of DNA fragments which migrates as a continuum on gels. A similar electrophoretic pattern is obtained with purified DNA digested by micrococcal nuclease. The presence of spermine (0.15 mM) and spermidine (0.5 mM) in preparative and incubation buffers decreases the rate of digestion of chromatin by endogenous endonuclease in situ approx. 10-fold, without affecting the size of the resulting DNA fragments. The rates of production of the smallest DNA fragments, monomer, dimer, and trimer, are nearly identical when high molecular weight DNA is present in excess, indicating that all of the chromatin multimers are equally susceptible to endogenous endonuclease. These observations points out the effects of various experimental conditions on the digestion of chromatin by nucleases.     

The structure of sub-nucleosomal particles. The octameric (H3/H4)4--125-base-pair-DNA complex

Chicken erythrocyte chromatin was depleted of histones H1, H5, H2A and H2B. The resulting (H3/H4)-containing chromatin was digested with micrococcal nuclease to yield monomer, dimer, trimer etc. units, irregularly spaced on the DNA, with even-number multimers being more prominent. Sucrose density gradient centrifugation separated monomers and dimers (7.7 S and 10.5 S). Sodium dodecyl sulphate gel electrophoresis and cross-linking indicated: the monomer contains 50-base-pair (bp), 60-bp and 70-bp DNA and the dimer 125-bp DNA; the monomer contains a tetramer and the dimer an octamer of H3 and H4. Partial association of monomer units to dimers inhibits structural studies of monomers. The internal structure of the dimer, i.e. and (H3/H4)4-125-bp-DNA particle, was studied using circular dichroism, thermal denaturation and nuclease digestion. Both micrococcal nuclease and DNase I digestion indicate that, unlike core particles, accessible sites occur in the centre of the particle and it is concluded that the (H3/H4)4-125-bp-DNA particle is not a 'pseudo-core particle' in which the 'extra' H3 and H4 replace H2A and H2B. It is proposed that the octamer particle is formed by the sliding together of two 'monomer' units, each containing the (H3/H4)2 tetramer and 70 bp of DNA. Excision of this dimer unit with micrococcal nuclease results in the loss of 10 readily digestible base pairs at each end, leaving 125 bp.     

Sequence specific cleavage of African green monkey alpha-satellite DNA by micrococcal nuclease

The sequence specificity of micrococcal nuclease complicates its use in experiments addressed to the still controversial issue of nucleosome phasing. In the case of alpha-satellite DNA containing chromatin from African green monkey (AGM) cells cleavage by micrococcal nuclease in the nucleus was reported to occur predominantly at only one location around position 126 of the satellite repeat unit (Musich et al. (1982) Proc. Natl. Acad. Sci. USA 79, 118-122). DNA control experiments conducted in the same study indicated the presence of many preferential cleavage sites for micrococcal nuclease on the 172 bp long alpha-satellite repeat unit. This difference was taken as evidence for a direct and simple phase relationship between the alpha-satellite DNA sequence and the position of the nucleosomes on the DNA. We have quantitatively analyzed the digestion products of the protein-free satellite monomer with micrococcal nuclease and found that 50% of all cuts occur at positions 123 and 132, 5% at position 79, and to a level of 1-3% at about 20 other positions. We also digested high molecular weight alpha-satellite DNA from AGM nuclei with micrococcal nuclease. Again cleavage occurred mostly at positions 123 and 132 of the satellite repeat unit. Thus digestion of free DNA yields results very similar to those reported by Musich et al. for the digestion of chromatin. Therefore no conclusions on a possible phase relationship can be drawn from the chromatin digestion experiments.     

Histone deacetylation is required for the maturation of newly replicated chromatin

The effects of inhibiting histone deacetylation on the maturation of newly replicated chromatin have been examined. HeLa cells were labeled with [3H]thymidine in the presence or absence of sodium butyrate; control experiments demonstrated that butyrate did not significantly inhibit DNA replication for at least 70 min. Like normal nascent chromatin, chromatin labeled for brief periods (0.5-1 min) in the presence of butyrate was more sensitive to digestion with DNase I and micrococcal nuclease than control bulk chromatin. However, chromatin replicated in butyrate did not mature as in normal replication, but instead retained approximately 50% of its heightened sensitivity to DNase I. Incubation of mature chromatin in butyrate for 1 h did not induce DNase I sensitivity: therefore, the presence of sodium butyrate was required during replication to preserve the increased digestibility of nascent chromatin DNA. In contrast, sodium butyrate did not inhibit or retard the maturation of newly replicated chromatin when assayed by micrococcal nuclease digestion, as determined by the following criteria: 1) digestion to acid solubility, 2) rate of conversion to mononucleosomes, 3) repeat length, and 4) presence of non-nucleosomal DNA. Consistent with the properties of chromatin replicated in butyrate, micrococcal nuclease also did not preferentially attack the internucleosomal linkers of chromatin regions acetylated in vivo. The observation of a novel chromatin replication intermediate, which is highly sensitive to DNase I but possesses normal resistance to micrococcal nuclease, suggests that nucleosome assembly and histone deacetylation are not obligatorily coordinated. Thus, while deacetylation is required for chromatin maturation, histone acetylation apparently affects chromatin organization at a level distinct from that of core particle or linker, possibly by altering higher order structure.     

Comparison of the subunit organization of early and late replicating chromatin

A comparison was made of the subunit organization of chromatin from regions of the genome with different metaphase chromosome banding characteristics by analyzing the accessibility of early and late replicating DNA in synchronized Chinese hamster ovary cells to digestion with staphylococcal nuclease. Three measures of nuclease susceptibility were employed: (1) the release of acid-soluble material; (2) a digestion index, P, which corresponds to the proportion of internucleosome segments which experienced at least one cleavage event; and (3) the size distribution of DNA fragments isolated from digested chromatin. Little or no difference was observed in the initial rates with which nuclease converted early and late replicating chromatin to acid-soluble material, although the initial digestion rates varied with time of cell collection in the cycle (metaphase > G1 mid-S > late-S or G2). Measurements of the digestion indices of material isolated from interphase cells suggested that initial cleavage events were more rapid in early replicating chromatin than in late replicating chromatin. In contrast, electrophoretic analysis revealed that oligomer DNA fragments from early labelled metaphase chromatin were slightly larger than corresponding fragments from late labelled metaphase chromatin. The size distribution of DNA in submonomer fragments obtained from extensively digested chromatin appeared to be identical regardless of the timing of replication or cell collection. Those small differences in chromatin digestibility that were observed may reflect subtle variations in the accessibility of internucleosome regions or perhaps in the higher-order arrangement of nucleosomes. However, no gross variation in accessibility to staphylococcal nuclease digestion was observed in chromatin localized to metaphase chromosome regions with vastly different cytological staining properties.     

Chromatin assembly in isolated mammalian nuclei.

Cellular DNA replication was stimulated in confluent monolayers of CV-1 monkey kidney cells following infection with SV40. Nuclei were isolated from CV-1 cells labeled with [3H]thymidine and then incubated in the presence of [alpha-32P]deoxyribonucleoside triphosphates under conditions that support DNA replication. To determine whether or not the cellular DNA synthesized in vitro was assembled into nucleosomes the DNA was digested in situ with either micrococcal nuclease or pancreatic DNase I, and the products were examined by electrophoretic and sedimentation analysis. The distribution of DNA fragment lengths on agarose gels following micrococcal nuclease digestion was more heterogeneous for newly replicated than for the bulk of the DNA. Nonetheless, the state of cellular DNA synthesized in vitro (32P-labeled) was found to be identical with that of the DNA in the bulk of the chromatin (3H-labeled) by the following criteria: (i) The extent of protection against digestion by micrococcal nuclease of DNase I. (ii) The size of the nucleosomes (180 base pairs) and core particles (145 base pairs). (iii) The number and sizes of DNA fragments produced by micrococcal nuclease in a limit digest. (iv) The sedimentation behavior on neutral sucrose gradients of nucleoprotein particles released by micrococcal nuclease. (v) The number and sizes of DNA fragments produced by DNase I digestion. These results demonstrate that cellular DNA replicated in isolated nuclei is organized into typical nucleosomes. Consequently, subcellular systems can be used to study the relationship between DNA replication and the assembly of chromatin under physiological conditions.     

Nucleosome assembly capacity of nuclear lysate after nuclease digestion

Limited digestion of lymphocyte nuclei with micrococcal nuclease degrades the nuclear DNA and results in a resistant plateau of about 50% of the original DNA. During the course of the nuclease cleavage as more and more DNA becomes acid-soluble an increasing amount of core histone is released from the disintegrated chromatin indicating that a part of nucleosomal protected DNA is degraded. These free histones appeared not to be different from those arising from resistant chromatin fragments. The released histones are in a native state which allows the exogenous DNA to be converted into nucleoprotein complexes which appear to exhibit a typical nucleosomal structure as tested by several criteria.     

Ion competition and micrococcal nuclease digestion studies of spermidine-condensed calf thymus DNA: Evidence for torus organization by circumferential DNA wrapping

Spermidine-condensed calf thymus DNA structures have been studied by ion competition using a sedimentation assay and by micrococcal nuclease digestion. Competitor ions Mg²⁺, Ca²⁺ and putrescine²⁺ show specific ion effects; but all three appear to affect the DNA condensation-decondensation equilibrium caused by spermidine³⁺ in a qualitatively similar manner, suggesting the spermidine³⁺-DNA interaction is largely electrostatic. Our data show a hysteresis in condensation and decondensation transition directions. We interpret this in terms of a kinetic block in the condensation direction with decondensation representing the equilibrium state of the system. These results agree with results obtained from related systems using different measurement techniques. Micrococcal nuclease digestion of spermidine-condensed calf thymus DNA produces broad but discrete bands in gel electrophoresis experiments. At least two bands determined to be 760 ± 87 bp and 1355 ± 135 bp, possess the size ratio 1:1.8 ± 0.4 consistent with their forming the monomer and dimer fragments of an arithmetic band series. We rationalize this result in terms of a localized micrococcal nuclease cleavage model of circumferentially-wrapped DNA toruses proposed previously by Marx, K.A. and Reynolds, T.C. (Proc. Natl. Acad. Sci. (1982) 79, 6484–6488). The arithmetic series monomer band (760 ± 87 bp), corresponding to wrapping B̄ DNA once circumferentially about the torus, is in agreement with the electron microscopic measurements of hydrated calf thymus DNA torus circumferences presented by Marx, K.A. and Ruben, G.C. (Nucleic Acids Res. (1983) 11, 1839–1853).     

Analysis of DNA double- and single-strand breaks by two dimensional electrophoresis: action of micrococcal nuclease on chromatin and DNA, and degradation in vivo of lens fiber chromatin.

We describe a novel system for two dimensional electrophoresis at neutral and alkaline pH for determining the double-stranded and single-stranded lengths of DNA. With this system we analysed the mode of micrococcal nuclease digestion of DNA in cellular and SV40 viral chromatin and of supercoiled SV40 DNA. The enzyme reaction occurred in two steps : the enzyme first introduced single-strand breaks, then converted these to double-strand breaks by an adjacent cleavage on the opposite strand. Digestion of cellular chromatin DNA occurred by a similar mechanism. Chromatin fragments produced by limited micrococcal nuclease action contained many single-strand breaks, which may be important when this method is used to prepare chromatin fragments for biochemical and biophysical studies. Nucleosome monomer to tetramer produced at later stages of digestion contained few if any single-strand breaks.     

Nuclease-sensitivity of methylated DNA as a probe for chromatin reconstitution by genotoxicants

DNA in Chinese hamster ovary cells was labeled with [¹⁴C]thymidine and [methyl- ³H]-1-methionine in culture, and their nuclei were digested with micrococcal nuclease. Not until 10 percent of bulk DNA was digested did methylated DNA appear in the acid-soluble fraction. When these cells were exposed to UV-radiation, alkylating agents and intercalating agents in culture, the resistance of methylated DNA to digestion by the nuclease was largely or completely eliminated. The change in the sensitivity of methylated DNA to the nuclease indicates a conformational change in chromatin induced by the genotoxicants.     

DNA-histone interaction in the vicinity of replication points.

Chromatin replication was studied in isolated nuclei from Concanavalin A activated lymphocytes. Digestion with micrococcal nuclease revealed that the resistant fraction of in vitro replicated DNA is associated with nucleosomes. Earlier experiments had shown that the nuclease resistant fraction of nascent DNA is composed of fragments which are shorter than the nuclease resistant fragments of bulk DNA. In this communication we demonstrate that the short fragments of nascent DNA are differently bound to nucleosome like structures compared to bulk DNA. At 0.5 M NaCl a fraction of pulse labeled labeled DNA is released from these structures and appears as free double stranded DNA of about 140 base pair length (5S DNA) while the 185 pair fragments of mature replicated DNA remain attached to nucleosomes under these conditions. The experiments may indicate that the interaction of a fraction of replicating DNA with histones differs from that of bulk DNA.     

Micrococcal nuclease digestion of nuclei reveals extended nucleosome ladders having anomalous DNA lengths for chromatin assembled on non-replicating plasmids in transfected cells.

The chromatin structures of a variety of plasmids and plasmid constructions, transiently transfected into mouse Ltk- cells using the DEAE-dextran procedure, were studied by micrococcal nuclease digestion of nuclei and Southern hybridization. Although regularly arranged nucleosome-like particles clearly were formed on the transfected DNA, the nucleosome ladders, in some cases with 13-14 bands, were anomalous. Most often, a ladder of DNA fragments with lengths of approximately 300, 500, 700, 900 bp, etc. was generated. In contrast, typical 180-190 bp multiples were generated from bulk cellular or endogenous beta-actin gene chromatin. Very similar results were obtained with all DNA's transfected, and in a variety of cell lines, provided that plasmid replication did not occur. Additionally, after digestion of nuclei, about 90% of the chromatin fragments that contained transfected DNA sequences could not be solubilized at low ionic strength, in contrast with bulk cellular chromatin, suggesting association with nuclear structures or nuclear matrix. The remaining 10% of transfected DNA sequences, arising from soluble chromatin fragments, generated a typical nucleosome ladder. These results are consistent with the idea that assembly of atypical chromatin structures might be induced by proximity to elements of the nuclear pore complex or by nuclear compartmentalization.     

cis-diamminedichloroplatinum (II) modified chromatin and nucleosomal core particle

The influence of cis-diamminedichloroplatinum (II) (cis-DDP) binding to chromatin in chicken erythrocyte nuclei and the nucleosomal core particle is investigated. The cis-DDP modifications alter DNA-protein interactions associated with the higher order structure of chromatin to significantly inhibit the rate of micrococcal nuclease digestion and alter the digestion profile. However, cis-DDP modification of core particle has little effect on the digestion rate and the relative distribution of DNA fragments produced by microccocal nuclease digestion. Analysis of the monomer DNA fragments derived from the digestion of modified nuclei suggests that cis-DDP binding does not significantly disrupt the DNA structure within the core particle, with its major influence being on the internucleosomal DNA. Together these findings suggest that cis-DDP may preferentially bind to the internucleosomal region and/or that the formation of the intrastrand cross-link involving adjacent guanines exhibits a preference for the linker region. Sucrose gradient profiles of the modified nucleoprotein complexes further confirm that the digestion profile for micrococcal nuclease is altered by cis-DDP binding and that the greatest changes occur at the initial stages of digestion. The covalent cross-links within bulk chromatin fix a sub-population of subnucleosomal and nucleosomal products, which are released only after reversal by NaCN treatment. Coupled with our previous findings, it appears that this cis-DDP mediated cross-linking network is primarily associated with protein-protein crosslinks of the low mobility group (LMG) proteins.     

Ion competition and micrococcal nuclease digestion studies of spermidine-condensed calf thymus DNA. Evidence for torus organization by circumferential DNA wrapping

Spermidine-condensed calf thymus DNA structures have been studied by ion competition using a sedimentation assay and by micrococcal nuclease digestion. Competitor ions Mg2+, Ca2+ and putrescine2+ show specific ion effects; but all three appear to affect the DNA condensation-decondensation equilibrium caused by spermidine3+ in a qualitatively similar manner, suggesting the spermidine3+-DNA interaction is largely electrostatic. Our data show a hysteresis in condensation and decondensation transition directions. We interpret this in terms of a kinetic block in the condensation direction with decondensation representing the equilibrium state of the system. These results agree with results obtained from related systems using different measurement techniques. Micrococcal nuclease digestion of spermidine-condensed calf thymus DNA produces broad but discrete bands in gel electrophoresis experiments. At least two bands determined to be 760 +/- 87 bp and 1355 +/- 135 bp, possess the size ratio 1:1.8 +/- 0.4 consistent with their forming the monomer and dimer fragments of an arithmetic band series. We rationalize this result in terms of a localized micrococcal nuclease cleavage model of circumferentially-wrapped DNA toruses proposed previously by Marx, K.A. and Reynolds, T.C. (Proc. Natl. Acad. Sci. (1982) 79, 6484-6488). The arithmetic series monomer band (760 +/- 87 bp), corresponding to wrapping B DNA once circumferentially about the torus, is in agreement with the electron microscopic measurements of hydrated calf thymus DNA torus circumferences presented by Marx, K.A. and Ruben, G.C. (Nucleic Acids Res. (1983) 11, 1839-1853).