Selection of transgenic flax plants is facilitated by spectinomycin

Regeneration of transformed flax shoots after inoculation withAgrobacterium tumefaciens carrying a binary vector with either a neomycin phosphotransferase (nptII) gene and a -glucuronidase (GUS) reporter gene or a spectinomycin resistance gene was examined. Hypocotyls from 4-day-old seedlings were inoculated with either of the twoA. tumefaciens strains. Selection and regeneration were achieved on a medium containing 0.1 M thidiazuron, 0.01 M napthalene acetic acid, 100 mgl^–1 kanamycin sulphate or spectinomycin sulphate and 300 mgl^–1 cefotaxime. Use of different neomycins for the selection of transformed tissues did select transformed calli but not transformed shoots either directly or via a callus phase. Selection based on spectinomycin resistance allowed the growth of transformed shoots. Transgenic shoots were rooted on a medium containing 100 mgl^–1 spectinomycin sulphate. Integration of the spectinomycin resistance gene into the flax genome was confirmed by Southern blot hybridizations and spectinomycin resistance was shown to be inherited as a dominant Mendeliant trait. Therefore, spectinomycin resistance is more suitable for genetic engineering of flax than aminoglycoside resistance.     

Mutational spectrum induced in Saccharomyces cerevisiae by the carcinogen N-2-acetylaminofluorene

The spectrum of mutations induced by the carcinogen N-2-acetylaminofluorene (AAF) was analysed in Saccharomyces cerevisiae using a forward mutation assay, namely the inactivation of the URA3 gene. The URA3 gene, carried on a yeast/bacterial shuttle vector, was randomly modified in vitro using N-acetoxy-N-2-acetylaminofluorene (N-AcO-AAF) as a model reactive metabolite of the carcinogen AAF. The binding spectrum of AAF to the URA3 gene was determined and found to be essentially random, as all guanine residues reacted about equally well with N-AcO-AAF. Independent Ura^– mutants were selected in vivo after transformation of the modified plasmid into a ura3 yeast strain. Plasmid survival decreased as a function of AAF modification, leading to one lethal hit (37% relative survival) for an average of 50 AAF adducts per plasmid molecule. At this level of modification the mutation frequency was equal to 70 × 10^–4, i.e. 50-fold above the background mutation frequency. UV irradiation of the yeast cells did not further stimulate the mutagenic response, indicating the lack of an SOS-like mutagenic response in yeast. Sequence analysis of the URA3 mutants revealed 48% frameshifts, 44% base substitutions and 8 % complex events. While most base substitutions (74%) were found to be targeted at G residues where AAF is known to form covalent C8 adducts, frameshift mutations were observed at GC base pairs in only 24% of cases. Indeed, more than 60% of frameshift events occurred at sequences such as 5-(A/T)_nG-3 where a short (n = 2 or 3) monotonous run of As or Ts is located on the 5' side of a guanine residue. We refer to these mutations as semi-targeted events and present a potential mechanism that explains their occurrence.     

Organization and function of cytochromeb and ubiquinone in the cristae membrane of beef heart mitochondria

The arrangement and function of the redox centers of the mammalianbc ₁ complex is described on the basis of structural data derived from amino acid sequence studies and secondary structure predictions and on the basis of functional studies (i.e., EPR data, inhibitor studies, and kinetic experiments). Two ubiquinone reaction centers do exist—a QH₂ oxidation center situated at the outer, cytosolic surface of the cristae membrane (Q₀ center), and a Q reduction center (Q _i center) situated more to the inner surface of the cristae membrane. The Q₀ center is formed by theb-566 domain of cytochromeb, the FeS protein, and maybe an additional small subunit, whereas the Q _i center is formed by theb-562 domain of cytochromeb and presumably the 13.4kDa protein (QP-C). The Q binding proteins are proposed to be protein subunits of the Q reaction centers of various multiprotein complexes. The path of electron flow branches at the Q₀ center, half of the electrons flowing via the high-potential cytochrome chain to oxygen and half of the electrons cycling back into the Q pool via the cytochromeb path connecting the two Q reaction centers. During oxidation of QH₂, 2H⁺ are released to the cytosolic space and during reduction of Q, 2H⁺ are taken up from the matrix side, resulting in a net transport across the membrane of 2H⁺ per e^– flown from QH₂ to cytochromec, the H⁺ being transported across the membrane as H (H⁺ + e^–) by the mobile carrier Q. The authors correct their earlier view of cytochromeb functioning as a H⁺ pump, proposing that the redox-linkedpK changes of the acidic groups of cytochromeb are involved in the protonation/deprotonation processes taking place during the reduction and oxidation of Q. The reviewers stress that cytochromeb is in equilibrium with the Q pool via the Q _i center, but not via the Q₀ center. Their view of the mechanisms taking place at the reductase is a Q cycle linked to a Q-pool where cytochromeb is acting as an electron pump.     

The bacterial phleomycin resistance geneble as a dominant selectable marker inChlamydomonas

A chimeric gene composed of the coding sequence of theble gene fromStreptoalloteichus hindustanus fused to the 5 and 3 untranslated regions of theChlamydomonas reinhardtii nuclear geneRBCS2 has been constructed. Introduction of this chimeric gene into the nuclear genome ofC. reinhardtii by co-transformation with theARG7 marker yields Arg⁺ transformants of which approximately 80% possess theble gene. Of these co-transformants, approximately 3% display a phleomycin-resistant (Pm^R) phenotype. Western blot analysis using antibodies against theble gene product confirms the presence of the protein in the Pm^R transformants and genetic analysis demonstrates the co-segregation of theble gene with the phenotype in progeny arising from the mating of a Pm^R transformant to wild-type strains. Direct selection of Pm^R transformants was achieved by allowing an 18-h period for recovery and growth of transformed cells prior to selection. This work represents the first demonstration of stable expression and inheritance of a foreign gene in the nuclear genome ofC. reinhardtii and provides a useful dominant marker for nuclear transformation.     

Over-production of the D1 protein of photosystem II reaction centre in the cyanobacterium Synechococcus sp. PCC 7942

The unicellular cyanobacterium Synechococcus sp. PCC 7942 has three psbA genes encoding two different forms of the photosystem II reaction centre protein D1 (D1:1 and D1:2). The level of expression of these psbA genes and the synthesis of D1:1 and D1:2 are strongly regulated under varying light conditions. In order to better understand the regulatory mechanisms underlying these processes, we have constructed a strain of Synechococcus sp. PCC 7942 capable of over-producing psbA mRNA and D1 protein. In this study, we describe the over-expression of D1:1 using a tac-hybrid promoter in front of the psbAI gene in combination with lacI ^Q repressor system. Over-production of D1:1 was induced by growing cells for 12 h at 50 mol photons m^-2 s^-1 in the presence of 40 or 80 g/ml IPTG. The amount of psbAI mRNA and that of D1:1 protein in cells grown with IPTG was three times and two times higher, respectively. A higher concentration of IPTG (i.e., 150 g/ml) did not further increase the production of the psbAI message or D1:1. The over-production of D1:1 caused a decrease in the level of D1:2 synthesised, resulting in most PSII reaction centres containing D1:1. However, the over-production of D1:1 had no effect on the pigment composition (chlorophyll a or phycocyanin/number of cells) or the light-saturated rate of photosynthesis. This and the fact that the total amounts of D1 and D2 proteins were not affected by IPTG suggest that the number of PSII centres within the membranes remained unchanged. From these results, we conclude that expression of psbAI can be regulated by using the tac promoter and lacI ^Q system. However, the accumulation of D1:1 protein into the membrane is regulated by the number of PSII centres.     

A Ser–Gly substitution in plastid-encoded photosystem II D1 protein is responsible for atrazine resistance in foxtail millet (<Emphasis Type="Italic">Setaria italica</Emphasis>)

A foxtail millet (Setaria italica L. Beauv.) line resistant to atrazine was obtained through interspecific hybridization between wild S. viridis L. Beauv. and cultivated S. italica. The resistance was proved to be controlled by a chloroplast-inherited gene and it has further been utilized in foxtail millet production. However, the sequence information of the putative atrazine resistance gene, psbA in foxtail millet’s chloroplast genome encoding photosystem II D1 protein (32 kDa thylakoid membrane protein) (photosystem Q_B protein) and the mutation site responsible for the resistance are not known. In this paper the psbA sequences of six atrazine susceptible/resistant foxtail millet varieties were obtained and compared. The results indicated that there was only one amino acid difference between susceptible and resistance gene, resulting from a single base substitution. It was concluded that a mutant allele of photosystem II protein D1 encoding a Gly residue instead of a Ser residue at position 264 is a major gene of resistance to atrazine. Moreover, the phylogenetic tree based on the psbA coding region of thirty-five plant species was carried out. The phylogenetic relationship between S. italica and other plants and the related evolutionary issues were discussed and it was suggested that psbA sequences could be used in phylogenetic studies in plants. Xiaoping Jia and Jincheng Yuan have equal contribution.     

The IleL229 → Met mutation impairs the quinone binding to the QB-pocket in reaction centers of Rhodobacter sphaeroides

A spontaneous mutant (R/89) of photosynthetic purple bacterium Rhodobacter sphaeroides R-26 was selected for resistance to 200 M atrazin. It showed increased resistance to interquinone electron transfer inhibitors of o-phenanthroline (resistance factor, RF=20) in UQ_o reconstituted isolated reaction centers and terbutryne in reaction centers (RF=55) and in chromatophores (RF=85). The amino acid sequence of the Q_B binding protein of the photosynthetic reaction center (the L subunit) was determined by sequencing the corresponding pufL gene and a single mutation was found (Ile^L229 Met). The changed amino acid of the mutant strain is in van der Waals contact with the secondary quinone Q_B. The binding and redox properties of Q_B in the mutant were characterized by kinetic (charge recombination) and multiple turnover (cytochrome oxidation and semiquinone oscillation) assays of the reaction center. The free energy for stabilization of Q_AQ_B ^– with respect to Q_A ^–Q_B was G_AB=–60 meV and 0 meV in reaction centers and G_AB=–85 meV and –46 meV in chromatophores of R-26 and R/89 strains at pH 8, respectively. The dissociation constants of the quinone UQ_o and semiquinone UQ_o ^– in reaction centers from R-26 and R/89 showed significant and different pH dependence. The observed changes in binding and redox properties of quinones are interpreted in terms of differential effects (electrostatics and mesomerism) of mutation on the oxidized and reduced states of Q_B.Abbreviations BChl bacteriochlorophyll - Ile isoleucine - Met methionin - P primary donor - Q_A primary quinone acceptor - Q_B secondary quinone acceptor - RC reaction center protein - UQ_o 2,3-dimethoxy-5-methyl benzoquinone - UQ₁₀ ubiquinone 50 This work is dedicated to the memory of Randall Ross Stein (1954–1994) and is, in a small way, a testament to the impact which Randy's ideas have had on the development of the field of competitive herbicide binding.     

The suppressor of Hairy-wing binding region is required for gypsy mutagenesis

Summary We undertook a deletional analysis of the gypsy retrotransposon in order to determine which sequences of the element are required for its mutagenic effect. We show that a phenotype indistinguishable from that ofy ² flies can be generated by transformingy ^– flies with a construct containing theyellow gene and a gypsy element located at the same insertion site inyellow as found iny ² flies. When flies are transformed with similar constructs in which increasing amounts of the 5 transcribed untranslated region of gypsy have been removed, either a partialy ² revertant or a completely revertant phenotype is obtained. These results yield direct proof that the region of gypsy to which thesu(Hw) protein binds is required for the generation of mutant phenotypes by this retrotransposon.     

Maintenance and killing efficiency of conditional lethal constructs inPseudomonas putida

Summary Conditional lethal (suicidal) genetic constructs were designed and employed in strains of Pseudomonads as models for containment of geneticallyengineerd microbes that may be deliberately released into the environment. A strain ofPseudomonas putida was formed with a suicide vector designated pBAP24h that was constructed by cloning the host killing gene (hok) into the RSF1010 plasmid pVDtac24 and placing it under the control of thetac promoter. Afterhok induction inP. putida only 40% of surviving cells continued to bear thehok sequences within 4 h of induction; in contrast, 100% of the cells in uninduced controls borehok. A few survivors that demonstrated resistance tohok-induced killing developed inP. putida, which may have been due to a mutation or physiological adaptation that rendered the membrane resistant tohok. Conditional lethal strains ofP. putida also were formed by insertinggef (a chromosomal homolog ofhok) under the control of thetac promoter into the chromosome using a transposon. Constructs with chromosomalgef, as well as an RK2-derived plasmid construct containinggef, were only marginally more stable than thehok constructs; they were effective in killingP. putida when induced and within 2 h post-induction killing from eithergef construct resulted in a 10³–10⁵-fold reduction in viable cell count compared to uninduced controls.     

Effect of some heavy metal ions on copper-induced metallothionein synthesis in the yeast Saccharomyces cerevisiae

Copper-induced metallothionein (MT) synthesis in Saccharomyces cerevisiae was investigated in order to associate this exclusively with Cu²⁺ in vivo, when cultured in nutrient medium containing other heavy metal ions. Expression of the CUP1 promoter/lacZ fusion gene was inhibited by all heavy metal ions tested, especially Cd²⁺ and Mn²⁺. By adding Cd²⁺ and Mn²⁺ at 10 M concentration, the -galactosidase activity decreased by about 80% and 50% of the maximum induction observed with 1 mM CuSO₄, respectively. Furthermore, cell growth was markedly inhibited by combinations of 1 mM-Cu²⁺ and 1 M-Cd²⁺. Therefore, the yeast S. cerevisiae could not rely on MT synthesis as one of the copper-resistance mechanisms, when grown in a Cd²⁺ environment. In contrast, the presence of Mn²⁺ in the nutrient medium showed alleviation rather than growth inhibition by high concentrations of Cu²⁺. The recovery from growth inhibition by Mn²⁺ was due to decreased Cu²⁺ accumulation. Inhibitory concentrations of Co²⁺, Ni²⁺ and Zn²⁺ on expression of the CUP1p/lacZ fusion gene were at least one order of magnitude higher than that of Cd²⁺ and Mn²⁺. These results are discussed in relation to Cu²⁺ transport and Cu-induced MT synthesis in the copper-resistance mechanism of the yeast S. cerevisiae.     

Nucleotide sequence of the chloroplast gene responsible for triazine resistance in canola

Summary The nucleotide sequence for the psbA gene from a triazine resistant cultivar of B. napus (cv Triton) has been determined. This gene encodes an open reading frame of 353 amino acids that is highly homologous to other higher plant psbA genes at both the nucleotide and amino acid levels. As has been found for other triazine resistant psbA genes, the Triton psbA contains an A to G nucleotide change which results in a serine to glycine amino acid substitution at position 264. The B. napus psbA gene also has a G insertion at position –9 resulting in a ribosome binding site sequence (AGGA) just before the initial methionine and suggesting that the entire open reading frame is translated. A large (72 bp) insertion is also found upstream of the B. napus psbA gene which resembles a similar insertion in the mustard psbA. The uncloneable nature of the entire gene is further investigated through reconstruction experiments and the implications discussed.     

The sequence of a pea vicilin gene and its expression in transgenic tobacco plants

A 5.5 kb Eco RI fragment containing a vicilin gene was selected from a Pisum sativum genomic library, and the protein-coding region and adjacent 5 and 3 regions were sequenced. A DNA construction comprising this 5.5 kb fragment together with a gene for neomycin phosphotransferase II was stably introduced into tobacco using an Agrobacterium tumefaciens binary vector, and the fidelity of expression of the pea vicilin gene in its new host was studied. The seeds of eight transgenic tobacco plants showed a sixteen-fold range in the level of accumulated pea vicilin. The level of accumulation of vicilin protein and mRNA correlated with the number of integrated copies of the vicilin gene. Pea vicilin was confined to the seeds of transgenic tobacco. Using immunogold labelling, vicilin was detected in protein bodies of eight out of ten embryos (axes plus cotyledons) and, at a much lower level, in two out of eleven endosperms. Pea vicilin was synthesized early in tobacco seed development; some molecules were cleaved as is the case in pea seeds, yielding a major parental component of M _r50000 together with a range of smaller polypeptides.     

Optimization of fed-batch culture by dynamic programming and regression analysis

Summary An algorithm for the optimization of a fermentation process was studied using the combination of dynamic programming and linear predictive procedure by regression analysis. It was applied to the fed-batch culture for glutamic acid production with ethanol feeding, the results of which proved that it was effective for the optimization problems.Nomenclature a _i , b _i , c _i , d _i ;i = 1 2 or 3: Partial regression coefficients derived by multiple regression analysis ( - ) - E : Euclid distance ( - ) - F : Volumetric feed rate (l/hr) - f : Defined by Eq. (13) (g-glutamic acid) - G : Concentration of glutamic acid (g/l) - G* : Aeration rate (l/hr) - J : Objective function defined by Eq. (12) (g-glutamic acid) - N : Number of stage ( - ) - P : Defined by Eq. (14) (g-glutamic acid) - Q : Metabolic activity of the culture, in this case Q = Q _{CO 2} (mole CO₂/g-cell hr) - S : Ethanol concentration in culture broth (g/l) - S _Q : Ethanol concentration in feed (g/l) - S* : Ethanol concentration in effluent gas (g/l) - t : Culture time (hr) - t _o : Initial culture time (hr) - V : Volume of culture broth (l) - X : Cell concentration (g/l) - X : State vector (X, S, Q, V) - : Rate of Q _{CO 2} change (mole CO₂/g-cell·h²) - : Specific growth rate of microorganisms (hr^–1) - : Specific production rate of glutamic acid (g-glutamic acid/g-cell·hr) - : Specific consumption rate of ethanol (g-ethanol/g-cell·hr) - : Standard deviation ( - ) This optimization procedure was presented in preliminary form at the 45th Annual Meeting of the Soc. of Chem. Engrs., Japan, Osaka, C105 (1980).     

Construction of a Chlamydomonas reinhardtii mutant with an intronless psbA gene

Efficient chloroplast transformation systems now available allow the manipulation of the evolutionarily highly conserved psbA gene in the eucaryotic organism Chlamydomonas reinhardtii. Two copies of this gene in the inverted repeat region of the chloroplast genome contain four large group I introns. To analyse possible functions of these introns and to generate a mutant for simplified psbA gene manipulations, a psbA cDNA fragment was introduced into a psbA deletion mutant using the biolistic transformation method. A transformant with no introns in the psbA gene has been obtained and represents the first example of the removal of a complete set of introns from a chloroplast gene. The newly generated strain is photosynthetically competent and contains no detectable recipient genome copies. The loss of all four introns appears to be phenotypically silent.     

Expression ofKlebsiella pneumoniae nif genes inProteus mirabilis

Self-transmissible plasmids carryinghis andnif genes fromKlebsiella pneumoniae have been introduced into threehis mutants ofProteus mirabilis: strains 5006-1, WR19 and WR20. Expression ofhis by the transconjugants was unequivocal, if slightly temperature-sensitive, but none was Nif⁺ when tested for acetylene reduction in anaerobic glucose medium using inocula from rich or glucose-minimal aerobic agar cultures. Succinate or pyruvate in place of glucose, low glucose, lower temperature or elevated Na₂MoO₄ did not allownif expression and no nitrogenase MoFe-protein peptide was detected immunologically after exposure to conditions in which diazotrophic enterobacteria, normal or genetically constructed, derepressnif.One strain,P. mirabilis WR19, carrying thehis nif Km^r plasmid pMF250 was examined in detail. Thenif activator genenifA was introduced on the plasmid pCK1. Such derivatives remained Nif^- when tested, after aerobic growth on rich agar media, with normal or low glucose, with succinate or with elevated Mo. However, pre-conditioning by aerobic growth on glucose-minimal agar led to subsequent anaerobic expression ofnif in glucose medium from pMF250 in WR19 carrying pCK1. NH ₄ ⁺ or proline could serve as N-source in the glucose-minimal agar. Maximum activity was about 5% of that ofK. pneumoniae in our assay conditions. Material cross-reacting with anti-serum to the nitrogenase MoFe protein was formed. Nitrogenase activity was not switched off by NH ₄ ⁺ .P. mirabilis WR19 (pCK1) showed NH ₄ ⁺ -constitutive temperature-sensitive kanamycin resistance (anif-related phenotype of this plasmid) in aerobic glucose minimal medium. Expression ofnif inP. mirabilis WR19 (pCK1, pMF250) was NH ₄ ⁺ -repressible despite the constitutivenifA character of pCK1 and introduction of thentrA ⁺ plasmid pMM17 did not alter this phenotype. However, pCK1 did not give rise to NH ₄ ⁺ -constitutive diazotrophy in the wild-typeK. pneumoniae M5al. A construct of WR19 carrying pMF250 and constitutiventrC plasmid (pMD45) remained Nif^- even after pre-growth on glucose-minimal media.We conclude (a) thatP. mirabilis forms a gene product functionally equivalent to that ofntrA inK. pneumoniae, (b) that it forms no functional equivalent of thentrC product in our growth conditions. The need for pre-conditioning on aerobic glucose media remains perplexing.Non-common abbreviation NFDM Nitrogen-free-Davis-Mingioli medium     

σs-dependent overexpression offtsZ in anEscherichia coli K-12 rpoB mutant that is resistant to the division inhibitors DicB and DicF RNA

TheEscherichia coli genesdicF anddicB encode division inhibitors, which prevent the synthesis and activity, respectively, of the essential division protein FtsZ. A mutation at the C-terminal end of the RNA polymerase subunit renders cells resistant to both inhibitors. In the mutant strain the level of theftsZ gene product is higher than in the wild type. Disruption ofrpoS, which encodes the stationary phase sigma factor ^S, lowers FtsZ protein levels in the mutant, and partially restores sensitivity to the inhibitors.     

Molecular characterization of the lysosomal acid phosphatase fromDrosophila melanogaster

InDrosophila, unlike humans, the lysosomal acid phosphatase (Acph-1) is a non-essential enzyme. It is also one of the most rapidly evolving gene-enzyme systems in the genus. In order to determine which parts of the enzyme are conserved and which parts are apparently under little functional constraint, we cloned the gene fromDrosophila melanogaster via a chromosomal walk. Fragments from the gene were used to recover an apparently full-length cDNA. The cDNA was subcloned into aDrosophila transformation vector where it was under the control of the 5 promoter sequence of thehsp-70 gene. Three independent transformants were obtained; in each, Acph-1 expression from the cDNA was constitutive and not dependent on heat shock, as determined by densitometric analyses of the allozymic forms of the enzyme. The pattern of expression indicates thehsp-70 and endogenousAcph-1 promoters act together in some, but not all, tissues. The sequence of the cDNA was determined using deletions made with exonuclease III, and primers deduced from the cDNA sequence were used to sequence the genomic clone. Five introns were found, and putative 5 up-stream regulatory sequences were identified. Amino acid sequence comparisons have revealed several highly conserved motifs betweenDrosophila Acph-1 and vertebrate lysosomal and prostatic acid phosphatases.     

Spontaneous mutations changing the raffinose metabolism ofLactobacillus plantarum

Lactobacillus plantarum ATCC 8014 grew poorly on raffinose agar plates, but large mutant colonies appeared in high frequency from a thin film of background growth. The -galactosidase and -galactosidase activities ofL. plantarum ATCC 8014 and a mutant strain were studied in static cultures and pH-controlled fermenter cultures. Both -galactosidase and -galactosidase production were inducible in the parental strain; the induction was not needed in the mutant. The -galactosidase activity of both strains was repressed by glucose but not by -methyl-D-glucoside. The mutant phenomenon might be an obstacle in connection to traditionalLactobacillus identification by means of carbohydrate fermentation.     

Isolation of Hem3 mutants from Candida albicans by sequential gene disruption

Summary Molecular methods for directed mutagenesis in Candida albicans have relied on a combination of gene disruption by transformation to inactivate one allele and UV-induced mitotic recombination or point mutation to produce lesions in the second allele. An alternate method which uses two sequential gene disruptions was developed and used to construct a C. albicans mutant defective in a gene essential for synthesizing tetrapyrrole (uroporphyrinogen I synthase). The Candida gene was cloned from a random library by complementation of the hem3 mutation in Saccharomyces cerevisiae. The complementing region was limited to a 2.0 kb fragment by subcloning and a BglII site was determined to be within an essential region. Linear fragments containing either the Candida URA3 or LEU2 gene inserted into the BglII site were used to disrupt both alleles of a leu2, ura3 mutant by sequential transformation. Ura⁺, Leu⁺ heme-requiring strains were recovered and identified as hem3 mutants by Southern hybridization, transformation to heme independence by the cloned gene, and enzyme assays.     

Elements of theS-gene complex VI. Mutations of the self-incompatibility gene,pseudo-compatibility and origin of new self-incompatibility alleles

Spontaneously occurring mutations of theS gene, involving both theS _I and theS _FI classes of alleles, were studied inNicotiana alata. The results showed that while almost all of the irradiation-induced mutants of theS gene requiredS-bearing duplication for their survival, usually in the form of a free fragment, most of the spontaneous mutants in the same species, surprisingly, did not have such a requirement. This difference has been attributed to the greater depth of mutations produced in response to the ionizing radiations, which necessitated complementation for the survival of the mutants. There is a possibility from the data that theS _FI class of alleles may have even less need for the duplication than the S_I class of alleles. Both pollen- and stylar-part mutations of theS gene were obtained, but the majority of the mutations were partial, producing less than half the normal complement of seeds per pod in the mutants. Complementation was observed in the style between a -part mutant alleleS _infF11 ^sup and a normal alleleS _F10, which was the other allele in the parental plant that produced the mutant. No complementation occurred with another normal unrelated alleleS ₂. This observation was similar to that previously recorded in the study of induced mutants inN. alata.In a cross where the two alleles of the pollen parent were both compatible the allele which was also a mutant had an advantage over the other, normal, allele. This suggests that in maize, where the occurrence of mutant forms of theS gene has been demonstrated, the preferential fertilization of ovules by pollen containing the B-chromosomes may be due to the presence of a mutant form of theS gene on the B-chromosome.Besides clear-cutS-gene mutants, there were others, showing mostly irregular, slight compatibility, which did not appear to be directly related to theS-gene mutation. In some of the progeny of certain of these mutants, partial or complete lack of the specificity of one or bothS alleles in the style was observed; in certain others all progeny were normal. This pseudo-compatibility is attributed to cytoplasmic mutations affecting the products of theS gene; however, the possibility of an effect of chance polygenic modifier combinations is not ruled out.Recent literature on theS-gene structure, mutational specificity ofS alleles, and genetic control of pseudo-compatibility is reviewed. The time ofS-gene action is discussed in relation to the mechanism of the generation of new self-incompatibility alleles.