Nuclear Gene Dosage Effects Upon the Expression of Maize Mitochondrial Genes

Corresponding author: James A. Birchler, Division of Biological Sciences, 117 Tucker Hall, University of Missouri, Columbia, MO 65211., birchlerj{at}missouri.edu (E-mail)

Communicating editor: V. L. CHANDLER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Each mitochondrion possesses a genome that encodes some of its own components. The nucleus encodes most of the mitochondrial proteins, including the polymerases and factors that regulate the expression of mitochondrial genes. Little is known about the number or location of these nuclear factors. B-A translocations were used to create dosage series for 14 different chromosome arms in maize plants with normal cytoplasm. The presence of one or more regulatory factors on a chromosome arm was indicated when variation of its dosage resulted in the alteration in the amount of a mitochondrial transcript. We used quantitative Northern analysis to assay the transcript levels of three mitochondrially encoded components of the cytochrome c oxidase complex (cox1, cox2, and cox3). Data for a nuclearly encoded component (cox5b) and for two mitochondrial genes that are unrelated to cytochrome c oxidase, ATP synthase -subunit and 18S rRNA, were also determined. Two tissues, embryo and endosperm, were compared and most effects were found to be tissue specific. Significantly, the array of dosage effects upon mitochondrial genes was similar to what had been previously found for nuclear genes. These results support the concept that although mitochondrial genes are prokaryotic in origin, their regulation has been extensively integrated into the eukaryotic cell.

MITOCHONDRIA are the cellular sites for many energy conversion processes. Each mitochondrion possesses a genome that encodes a portion of the components necessary for its own biogenesis and functioning. The nuclear genome codes for most of the mitochondrially localized proteins. Mitochondrial transcript levels appear to be subject to cell-specific regulation (e.g., YOUNG and HANSON 1987 ; SINGH and BROWN 1993 ; MONEGER et al. 1994 ; SMART et al. 1994 ; LI et al. 1996 ). Although nuclear genes are thought to control mitochondrial transcript levels by influencing transcription and RNA processing, few such genes have been identified. Mitochondrial transcription is accomplished by nuclearly encoded phage T7-like RNA polymerase and associated transcription factors (NEWTON et al. 1995 ; GRAY and LANG 1998 ).

Direct evidence of the interaction between nuclear and mitochondrial genes comes from the study of cytoplasmic male sterility (CMS). CMS is an economically important plant trait that facilitates the production of hybrid seed without manual emasculation of plants (reviewed by HANSON and FOLKERTS 1992 ; LEVINGS 1993 ). Mitochondrial CMS genes appear to be novel chimeric DNA sequences whose expression perturbs the development of pollen. Genes involved in restoring fertility (Rf genes) are nuclear and often act by reducing the abundance of CMS transcripts (reviewed by SCHNABLE and WISE 1998 ).

CMS can also result from divergent nuclear-cytoplasmic interactions. For example, a nuclear genome from the A619 maize inbred in combination with Zea perennis cytoplasm results in plants that are apparently normal except that they are male sterile (GRACEN and GROGAN 1974 ). Most other maize inbred lines possess dominant nuclear genes that restore male fertility (GRACEN and GROGAN 1974 ; LAUGHNAN and GABAY-LAUGHNAN 1983 ). Up to three nonallelic dominant genes have been identified that each independently restores fertility to Z. perennis cytoplasm (NEWTON and COE 1989 ).

CMS is not the only effect resulting from divergent nuclear-cytoplasmic interactions. A combination of any teosinte cytoplasm from the section Luxurians with the maize inbred W23 nuclear genome results in plants that are small, pale, and slow growing (ALLEN et al. 1989 ). This "teosinte-cytoplasm-associated miniature" (TCM) trait also results in reduced kernel size. Two dominant nonallelic nuclear genes have been identified that rectify the TCM trait and result in normal plants (ALLEN et al. 1989 ).

Plants with divergent nuclear-cytoplasmic combinations have also demonstrated nuclear-mitochondrial interactions at the molecular level. In Z. perennis and many maize plants with Z. perennis cytoplasms the most abundant cytochrome c oxidase subunit 2 (cox2) transcript is a 1.9-kb mRNA. If the maize nuclear genome possesses the dominant Modifier of cox2 transcript (Mct) gene, then a novel 1.5-kb transcript predominates and the 1.9-kb transcript level is reduced (COOPER et al. 1990 ). The novel transcript is due to the activation of a promoter sequence that had been previously undescribed (NEWTON et al. 1995 ). Thus, Mct is a nuclear factor that affects the size and abundance of cox2 mRNAs by determining transcription start sites.

The purpose of this study was to investigate the interaction of the nuclear genome with mitochondrial gene expression levels in maize with normal cytoplasm. B-A reciprocal translocations were used to create chromosomal dosage series for fourteen chromosome arms. These chromosomal dosage series were used to determine which chromosome arms, when varied in dosage, affected the level of expression of various mitochondrial genes. The basis of the chromosomal dosage effects is believed to be rate-limiting factors that modify gene expression (GUO and BIRCHLER 1994 ). The effect could be at the level of transcription or it could involve a post-transcriptional aspect of gene expression, such as RNA stability. The dosage effects of 14 chromosome arms were determined for RNA in both embryo and endosperm tissue of kernels 30 days postpollination.

Previous studies have indicated an extensive array of trans-acting dosage effects on nuclear genes (BIRCHLER and NEWTON 1981 ; GUO and BIRCHLER 1994 ). Mitochondrial genomes, having evolved from a prokaryotic progenitor, could differ significantly in regulatory processes, so it was of interest to determine whether the same types of effects were operative on organellar genes. We also examined the degree to which related mitochondrial genes are coordinately affected. Finally, we determined the effects upon the expression level of a nuclearly encoded mitochondrial protein. The effects upon the nuclear gene were compared to the effects upon the mitochondrial genes to determine whether they are subject to the same types of modulations.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Production and identification of aneuploids:
Dosage series were produced using B-A reciprocal translocations. B chromosomes are supernumerary and often nondisjoin at the second pollen mitosis. This results in one sperm of a pollen grain having two B chromosomes and the other having none (ROMAN 1947 ). A B-A translocation involves an exchange of part of the B chromosome with one of the 10 chromosomes that make up the haploid set of maize (i.e., an A chromosome). The translocation chromosome with the B centromere and an arm of an A chromosome generally behaves like the B chromosome and regularly nondisjoins at the second pollen mitosis (see BECKETT 1978 ). In that case one sperm has two copies of the translocated A chromosome arm and the other has none. This creates an opportunity to produce aneuploids. Fertilization of the egg with the sperm that possesses two B-A's yields individuals that are hyperploid for the A chromosome arm; fertilization with the deficient sperm results in progeny that are hypoploid for that same chromosome arm. Euploid progeny are also produced because some of the pollen grains possess sperm where the B-A's disjoined normally.

Fifteen different maize stocks, each possessing a different B-A translocation, were used to generate dosage series (Table 1). Each stock is named for the translocation present. Translocations are designated according to which chromosomes are involved. Maize chromosomes are numbered from 1 to 10, each possessing a long (L) and a short (S) arm. TB-1Sb involves a translocation between a B chromosome and the short arm of chromosome 1. Each translocation is also given an alphabetical or numerical suffix so that it can be distinguished from other translocations involving the same chromosomes. For example, we have included both TB-10L19 and TB-10L32. TB-10L19 has all of 10L attached to the B centromere, while TB-10L32 has only the most distal 26% of 10L (MAGUIRE 1985 ). To simplify our text we omitted the TB prefix because in each case it is understood that what has been manipulated was a translocation with a B chromosome.

Each translocation stock had the B-A chromosome marked with a dominant genetic marker, either an anthocyanin color gene or an autonomous transposable element (Table 1). Plants with marked B-A chromosomes were crossed onto an appropriate tester plant. Testers possessed either a recessive allele of a color marker gene or a reporter gene that produces colored spots in the presence of an autonomous transposable element (Table 1). Reporter genes are color factors that are nonfunctional due to the insertion of a nonautonomous transposable element. In the presence of a related autonomous element, the nonautonomous element sporadically excises from the gene, causing it to become functional.

The marker system allowed for the determination of chromosome arm dosage because nondisjunction of the B-A chromosome resulted in noncorrespondence of embryo and the endosperm phenotypes (BIRCHLER and ALFENITO 1993 ). When the sperm with two B-A's fertilized the egg, the embryo was hyperploid and colored (or spotted). These colored embryos had hypoploid colorless endosperm because the other sperm of the pollen grain had no B-A and this deficient sperm fertilized the central cell. Alternatively, when the sperm with two B-A's fertilized the central cell, the endosperm was hyperploid and colored and the embryo was hypoploid and colorless. Normal disjunction of the B-A at the second pollen mitosis resulted in both the embryo and endosperm being euploid and having color since each received one copy of the B-A.

Plants were grown and pollinations were made as described by NEUFFER 1994 . Kernels were harvested 30 days after pollination. Embryos and endosperms were classified, separated, and immediately frozen in liquid nitrogen. Tissues were stored at -80°.

RNA isolation:
RNA was extracted from frozen tissue. Typically 10–20 embryos (1 g) from the same ear with the same chromosome arm dosage were pooled. Similarly, 5–15 endosperms (2 g) were combined. Total RNA was isolated using a method similar to that used by CONE et al. 1986 . Tissue was ground to a fine powder in liquid nitrogen using a mortar and pestle. About 10 ml of lysis buffer was rapidly and vigorously added per gram of frozen ground tissue. The lysis buffer was 2% sodium dodecyl sulfate (SDS), 0.1 M sodium chloride (NaCl), 50 mM Tris-HCl (pH 7.4), 50 mM disodium ethylenediaminetetraacetate (EDTA; pH 8), and 200 µg/ml proteinase K. The supernatant was subjected to three extractions: the first two with phenol (pH 7.9):chloroform (1:1) and the last with pure chloroform. After each extraction the samples were centrifuged at 12,000 x g for 10 min. Total RNA was precipitated with LiCl (2 M final solution). The RNA was redissolved and reprecipitated with 2.5 volumes absolute ethanol and 0.1 volume 3 M sodium acetate (pH 5.4). The RNA was then redissolved in formamide and the concentration was determined by measuring the optical density of a sample at 260 nm.

Electrophoresis and blotting:
Samples were subjected to electrophoresis in formaldehyde-agarose (1.5%) gels. Each lane was loaded with 5 µg of total RNA. RNA was transferred to uncharged nylon membranes by capillary blotting. The RNA was immobilized on the membranes by UV cross-linking.

Methylene blue staining:
Blots were stained with methylene blue (HERRIN and SCHMIDT 1988 ). They were immersed in 5% acetic acid and gently agitated for 15 min. The acetic acid was replaced with a methylene blue staining solution (0.5 M sodium acetate, pH 5.4; 0.04% methylene blue) and allowed to agitate for another 10 min. Blots were rinsed in deionized distilled water. The moist blots were wrapped in clear plastic and scanned on a conventional flat bed scanner.

Hybridization:
Hybridization conditions were essentially the same as in BIRCHLER and HIEBERT 1989 for RNA antisense probes. The blots were prehybridized for 4 hr at 60°. A ³²P-radiolabeled probe was added and allowed to hybridize for an additional 20 hr at the same temperature. Except for the nuclear 26S rRNA, antisense RNA probes were prepared by in vitro transcription using T3, T7, or Sp6 RNA polymerase (Promega, Madison, WI). Probes for rRNA were synthesized by PCR using T3 and T7 primers. The hybridization solution was 50% formamide, 5x SSC (0.75 M NaCl; 0.075 M sodium citrate), 10x Denhardt's solution, 0.5% SDS, 10% dextran sulfate, and 0.2 mg/ml fragmented salmon sperm DNA. The blots were washed for 2 hr at 75° with four changes of wash solution. The wash solution was 0.2x SSC and 0.05% SDS. When a double-stranded DNA probe was used to detect rRNA, the prehybridization and hybridization temperature was 55°, the wash temperature was 65°, and no dextran sulfate was used in the hybridization buffer.

Probe sources:
All clones were inserted into plasmids that allowed for in vitro transcription (Table 2). The clone for cytochrome c oxidase subunit 1 (cox1) was obtained from E. KUZMIN (unpublished results), who cloned a 0.57-kb Pst1/ApaLI fragment from a maize CMS-S -library. This fragment covers the 5' portion of the cox1 gene. Part of the original clone for maize cox2 (FOX and LEAVER 1981 ) was subcloned into pBluescript KS+ by K. Yamato in the K. Newton laboratory. The original clones for cytochrome c oxidase subunit 3 (cox3; HIESEL et al. 1987 ) and mitochondrial 18S rRNA (rrn18; MULLIGAN et al. 1988 ) were subcloned into pBluescriptII KS+ by M. Guo in the J. Birchler laboratory. M. Guo also subcloned a 0.8-kb XhoI/EcoRI fragment of 26S rDNA into pBluescriptII KS+ from an original 9-kb clone (MESSING et al. 1984 ). The clone for cytochrome c oxidase subunit Vb (cox5b) provided by PIONEER HI-BRED INTL. (unpublished results) was an expressed sequence tag that was identified by its homology with cox5b of rice (KADOWAKI et al. 1996 ). The clone for ATP synthase -subunit (atpA; MULLIGAN et al. 1988 ) was provided in pBluescript KS+.

Data analysis:
Hybridized transcripts were quantitated using a Fuji (Stamford, CT) phosphorimager. Nuclear 26S rRNA was used as the gel-loading control. It was previously demonstrated that rRNA is a reliable loading control that shows no demonstrable fluctuation in expression level for the various chromosome arm dosage series (GUO and BIRCHLER 1994 ). The signal strength of the rRNA was obtained either from the methylene blue image or from subsequently probing the blot with a radiolabeled rRNA probe. The methylene blue images were quantitated using the same Fuji software that was used for the radiolabeled probes. The relative quantity of the probed RNA was determined from the ratio of the mRNA/rRNA signal strength.

For each chromosome arm there were three doses: hypoploid, euploid, and hyperploid. Three samples were made for each tissue at each dosage. Means and standard errors were calculated from the three samples. Because the essential question was how mRNA levels of an aneuploid differ from the euploid, the mean of the mRNA/rRNA value for each dose was divided by the mean of its respective euploid mRNA/rRNA signal. This resulted in all euploids having an expression level of one and any variation in the expression level of the aneuploid was evident as a deviation from one. Significant differences between aneuploid and euploid expression levels were determined by t-tests (P 0.05).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The expression levels of five mitochondrial genes were assayed. Three of them (cox1, cox2, and cox3) encode subunits of the cytochrome c oxidase complex of the mitochondrial electron transport chain. The two other mitochondrial genes encode proteins involved with the ATP synthase and ribosomal complexes (atpA and rrn18). We also assayed effects upon a nuclearly encoded component of the cytochrome c oxidase complex (cox5b).

Three genes showed a multiple banding pattern (Fig 1). In each case, one band was selected to obtain data for this study. The cox1 gene produced two transcripts. Our estimated sizes of 2.3 and 2.2 kb were very close to the published transcript sizes of 2.4 and 2.3 kb (ISAAC et al. 1985A ). The data presented in Fig 2 and Fig 3 are for the more abundant larger molecular weight transcript. The cox2 gene produces multiple transcripts ranging from 2.6 to 0.6 kb (FOX and LEAVER 1981 ). Although as many as 11 have been reported (LUPOLD et al. 1999A ), 5 bands showed prominently on our blots. Again, the data presented are for the highest-molecular-weight transcript, estimated on our blots as 2.7 kb in size. It was typically the most intense band. Although atpA had multiple bands, only the 2.5-kb band was consistently observed on all blots and it was used for this study. It is interesting to note that ISAAC et al. 1985B described a complex banding pattern with three major transcripts (2.5, 3.0, and 4.5 kb), while BRAUN and LEVINGS 1985 indicated two transcripts, a strong 2.6-kb band and a much weaker 5.0-kb band. Both of those studies used RNA from coleoptiles.

View larger version (37K): In this window In a new window Download PPT slide   Figure 1. RNA species with multiple transcripts. Three genes gave multiple transcripts: cox1, cox2, and atpA. Arrows indicate the particular transcripts used in this study. Note that while atpA gave multiple transcripts, only the 2.5-kb transcript was reliably represented on all blots. A double-headed arrow indicates this.

View larger version (98K): In this window In a new window Download PPT slide   Figure 2. The mRNA levels of six genes for 14 dosage series in embryo tissue. Each dosage series was composed of hypoploid (hypo), euploid (eup), and hyperploid (hyper) samples. The mRNA level for each member of a dosage series is shown relative to the euploid level (set at 1.00) along with its standard error (SE). The B-A translocation used to create each dosage series is listed on the left under B-A. Values that are significantly different (P 0.05, t-test) from the euploid are indicated by shaded boxes.

View larger version (98K): In this window In a new window Download PPT slide   Figure 3. The mRNA levels of six genes for 14 dosage series in endosperm tissue. The data are presented in the same manner as in Fig 2.

The data for the expression levels of all six genes are presented in Fig 2 for the embryo tissue and Fig 3 for the endosperm tissue. Although each dosage series is composed of three conditions (hypoploid, euploid, and hyperploid), statistical analysis was performed by pairwise comparisons of aneuploid to the euploid. A shaded box indicates aneuploid expression levels that were significantly different from the euploid (P 0.05 per t-test). In each pairwise comparison, the effect was considered positive if the sample with the higher dosage had an increased expression level, i.e., hypoploid less than euploid or euploid less than hyperploid. The effect was considered negative if the higher ploidy sample had a lower expression level, i.e., hypoploid greater than euploid or euploid greater than hyperploid.

There were eight combinations of effects that resulted from dosage series (Fig 4). In several cases both the hypoploid and hyperploid resulted in effects relative to the euploid that were consistent; i.e., both were positive or both were negative. These are called series effects. The effect of the 5Lb dosage series upon cox2 in the embryo is shown as an example of a positive series effect; i.e., the transcript level increases with the copy number of 5L. The effect of the 1La dosage series on cox5b in the embryo is shown as an example of a negative series effect. More rare were the mixed effects (Fig 4), where the hypoploid and hyperploid effects were opposite from each other. If both aneuploids had transcript levels that were less than the euploid, then the effect of hypoploidy was positive and the effect of hyperploidy was negative. If both aneuploids had transcript levels that were greater than the euploid, then the effect of hypoploidy was negative and the effect of hyperploidy was positive. The two types of mixed effects are exemplified by the pattern of 5Sc on atpA and 8Lc on cox2 in embryo tissue (Fig 4). More common were the cases where the effect occurs either in only the hypoploid or the hyperploid. Four examples, all from endosperm comparisons, are shown in Fig 4 to illustrate these cases.

View larger version (43K): In this window In a new window Download PPT slide   Figure 4. Examples of chromosomal effects. Each dosage series was composed of hypoploid (hypo), euploid (eup), and hyperploid (hyper) samples. Shaded boxes indicate significant differences from the euploid. Below the numerical values for each example are bar graph representations of the data. Below that the mRNAs and nuclear 26s rRNAs (loading control) from the Northern blots are shown. The gene, dosage series, and tissue of each example are listed at the bottom.

Dosage effects upon mitochondrial genes:
The array of effects upon the mitochondrial genes (Fig 2 and Fig 3, cox1, cox2, cox3, atpA, and rrn18) were generally similar to what was found in a previous study on chromosomal dosage on nuclearly encoded genes (GUO and BIRCHLER 1994 ). All chromosome arms had an effect upon one or more gene products but no chromosome arm affected all genes in all tissues. All genes were affected by more than one dosage series; no gene was affected by all dosage series. Each gene was affected by an average of seven dosage series per tissue. A notable variant from this average was rrn18 with half the number of dosage series that affected its expression level. These results are consistent with the idea that these effects are due to the action of individual factors distributed throughout the genome (GUO and BIRCHLER 1994 ).

The predominant effect was negative. Forty-two significant pairwise effects (i.e., hypoploid vs. euploid or euploid vs. hyperploid) were detected in the embryo tissue. Of these, 14 were positive and 28 were negative. There were 50 significant pairwise effects detected in the endosperm tissue. Of these, 21 were positive and 29 were negative. For the most part, effects were tissue specific; i.e., similar dosage effects upon a gene were unlikely to be found in both embryo and endosperm. Some exceptions were the effects of 1La and to a lesser extent the effects of 6Lc (Fig 2 and Fig 3). Comparison of the effects of 1La on the mitochondrial genes in embryo vs. endosperm yielded only two differences between embryo and endosperm. The dosage effects of 6Lc were less consistent except that, in all cases where effects were detected, they were negative.

Differences in the effects between the 2.3- and 2.2-kb cox1 transcripts:
Although the 2.3-kb cox1 transcript was selected for this survey, it was apparent that the 2.2-kb band was affected differently in several dosage series. The two cox1 transcripts result from two transcriptional start sites that are 95 bases apart (ISAAC et al. 1985A ). Unless the 5' regions contribute to a difference in RNA stability, any difference in transcript levels is due to a relative variation in transcription rate from each start site. The two transcripts were rather closely spaced on the Northern blots. Nevertheless, we were able to detect several instances where a dosage effect upon the 2.3-kb transcript was significantly different from the effect upon the 2.2-kb transcript. The effect of the 1La dosage series in embryo tissue is illustrated in Fig 5. The 2.3-kb band appeared to be subject to a stronger negative dosage effect than the 2.2-kb band. The black bars represent the 2.3-kb bands and the gray bars represent the 2.2-kb bands. The 2.3-kb expression levels are shown relative to the 2.3-kb euploid and the 2.2-kb expression levels are shown relative to the 2.2-kb euploid. The 1La dosage series yielded a similar result in the endosperm (Fig 6).

View larger version (63K): In this window In a new window Download PPT slide   Figure 5. Effect of 1La dosage series upon the 2.3- and 2.2-kb cox1 transcripts. The expression levels for the 2.3-kb transcripts (relative to the 2.3-kb euploid level) are shown as solid bars. The expression levels for the 2.2-kb transcripts (relative to the 2.2-kb euploid level) are shown as shaded bars. The mRNAs and rRNAs for each dosage are shown below.

View larger version (45K): In this window In a new window Download PPT slide   Figure 6. Comparison of 2.3- and 2.2-kb cox1 transcripts. Data are shown for those dosage series where the 2.3-/2.2-kb transcript signal ratio of an aneuploid was significantly different (P 0.05, t-test) from the euploid (indicated by asterisks). Shaded boxes indicate aneuploid expression levels that are significantly different from the euploid (P 0.05, t-test).

Thirteen aneuploids had a 2.3-/2.2-kb signal ratio that was significantly different from the euploid. Dosage data for these are presented in Fig 6. Shading indicates that the expression level of the aneuploid was significantly different (P 0.05, t-test) from the euploid. Asterisks indicate that the 2.3-/2.2-kb signal ratio of an aneuploid was significantly different (P 0.05, t-test) from the euploid 2.3-/2.2-kb signal ratio. In seven cases, the effect upon the 2.2-kb band was less or none (embryo 1La hypoploid and hyperploid, 3La hypoploid, 10L19 hypoploid; endosperm 1La hypoploid, 4Lb hypoploid, and 5Sc hypoploid). For five aneuploids, the effect upon the 2.2-kb band was greater (embryo 4Lb hypoploid, 5Sc hyperploid, 6Lc hypoploid; endosperm 1Sb hypoploid, and 4Sa hypoploid). The 5Sc hypoploid in the embryo was interesting because it resulted in a slight negative effect upon the 2.3-kb transcript and a slight positive effect upon the 2.2-kb band. The actual effects were likely somewhat stronger since the signals of the two closely spaced transcripts partially overlapped, reducing the differences.

Multiple transcripts were also produced by cox2 and usually atpA; in each case only one band was selected for this survey of chromosomal dosage effects (Fig 2 and Fig 3). As a rule, the additional transcripts responded to a dosage series in the same way as the selected transcript. There were several instances where it was apparent that one or more of the additional transcripts responded differently. Unlike cox1, no attempt was made to measure the other transcripts for cox2 and atpA because the exact nature of each band is unknown. It is likely that some of the differences in effects observed among the cox2 transcripts were due to differences in transcription because this gene has at least five separate promoters (LUPOLD et al. 1999A ). This could be particularly interesting because transcriptional control of cox2 appears to involve recombination of the mitochondrial genome, positioning various promoters upstream of the gene (LUPOLD et al. 1999B ).

Dosage series effects upon cox5b:
Effects upon cox5b (Fig 2 and Fig 3) were included for comparison. This component of the cytochrome c oxidase complex is encoded in the nucleus. Because the chromosomal location of cox5b is unknown in maize, it is possible that its copy number may have been varied in one of the dosage series. Transcript levels of cox5b were affected by a number of dosage series, but none of them displayed a structural gene dosage effect. A structural gene dosage effect would have been indicated if expression levels correlated to gene dosage; e.g., one dose resulted in an expression level that was 50% of two doses. Failure to detect a structural gene dosage effect does not eliminate the possibility that cox5b is located on one of the B-A's used in this study. It appears that in maize many genes are subject to dosage compensation in whole chromosome arm aneuploidies (GUO and BIRCHLER 1994 ). Dosage compensation occurs when a dosage-dependent negative modifier is varied concurrently with the structural gene (BIRCHLER 1981 ; BIRCHLER and NEWTON 1981 ).

In general, the extent and type of effects seen for cox5b appeared to be similar to the mitochondrial genes with one notable difference. Whereas mitochondrial genes mostly experienced negative effects in both embryo and endosperm, the effects upon cox5b in the two tissues were different. Embryo effects were mostly negative (five positive vs. eight negative) while endosperm effects were positive (five positive vs. zero negative). As with the mitochondrial genes, effects were generally tissue specific. Only 10L19 showed similar effects in both the embryo and the endosperm.

Dosage effects of TB-10L32 vs. TB-10L19:
The dosage series of TB-10L19 involved aneuploidy of the whole long arm of chromosome 10 (10L). For comparison we obtained data from a TB-10L32 dosage series, where only the distal 26% of 10L was varied in copy number (MAGUIRE 1985 ). Whereas 10L19 affected four genes in the embryo and three in the endosperm, 10L32 affected only one gene (cox5b) and this was in both tissues (Fig 7). These results indicate that the factor or factors involved with the dosage effects on cox1, cox3, and atpA are on the proximal 74% of 10L. The dosage effects on cox5b may be explained by a factor on the distal 26% of 10L.

View larger version (42K): In this window In a new window Download PPT slide   Figure 7. Effects of 10L19 and 10L32 upon cox5b in the embryo and endosperm. The data are presented in a similar manner as in Fig 2. The 10L19 dosage series involves all of 10L. The 10L32 dosage series involves only the distal 26% of 10L.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

This study establishes that multiple regions of the nuclear genome affect mitochondrial gene expression levels. A change in transcript abundance in a dosage series indicates the presence of one or more trans-acting factors on the affecting chromosome arm. GUO and BIRCHLER 1994 conducted a similar study of B-A dosage effects upon six nuclearly encoded genes. In that study each gene was affected by an average of six dosage series. This number is close to the average of seven for mitochondrial genes, especially considering that there was one less dosage series (9Lc) in the nuclear gene study. As with the mitochondrial genes, most of the effects upon nuclear genes were negative (37 positive vs. 53 negative). However, when embryo and endosperm effects were considered separately, the profile of effects differed. While embryo effects were predominantly negative (5 positive vs. 33 negative), endosperm effects were mostly positive (32 positive vs. 20 negative). The profile of effects upon the nuclearly coded cox5b appears to follow the trend of the other nuclear genes (Table 3). At least in the endosperm, the effects indicate that mitochondrial genes are subject to a different mode of dosage-sensitive effects than are nuclear genes, which may indicate a different mode of regulation.

The results demonstrate that there are numerous dosage-dependent nuclear factors that affect mitochondrial transcript levels; these modulations might result from a variety of mechanisms. Transcript abundance is affected not only by transcription and turnover rates, but also potentially by a change in mitochondrial genome copy number. This could be due to variation in the number of mitochondria per cell or a change in the copy number of genomes per mitochondrion. In the formation of the maize tapetum, cells can experience up to a 40-fold increase in the number of mitochondria per cell (LEE and WARMKE 1979 ). Differences in mitochondrial numbers have been described in other maize tissues (CLOWES and JUNIPER 1964 ; JUNIPER and CLOWES 1965 ). However, such copy number differences might be expected to result in a global change among all of the mitochondrial gene transcript levels. No strong global effects were apparent.

This result does not rule out the possibility that a dosage-sensitive nuclear factor could alter mitochondrial gene copy number. A global effect due to a change in mitochondrial genome number could be masked if a second factor on a chromosome arm had a specific effect upon one of the mitochondrial genes being assayed. Alternatively, changes in gene copy number may not have involved the whole mitochondrial genome but, instead, some portion of it. The organization of the maize mitochondrial genome is complex; subgenomes are formed through intramolecular recombination (FAURON et al. 1995 ). Differential amplification of a subgenome would change the copy number of only a subset of genes.

Mitochondrial ribosomal RNA (rrn18) was included in the study with the expectation that it would be relatively insensitive to modulation. Indeed, of all the genes assayed, rrn18 had the least number of effects and most of those were modest. If a particular aneuploidy affected a number of the mitochondrial genes including rrn18, it could have been an indication that mitochondrial genome copy number was being affected. Hyperploid 1Sb in endosperm had the same effect upon the RNA levels of all of the mitochondrial genes except cox2. Also in the endosperm, hyperploid 2Sa had negative effects upon all of the tested mitochondrial genes while hypoploid 2Sa had negative effects upon cox1, cox3, and atpA. However, the effects of 1Sb or 2Sa were not particularly strong. If any dosage series affected mitochondrial genome copy number, it did not appear to contribute greatly to the effects upon RNA levels.

The 6Lc dosage series was interesting in that it may have indicated a different trend. Although it had no effect upon rrn18, the 6Lc hypoploid negatively affected all of the other tested mitochondrial genes. While the 2.3-kb cox1 transcript was not affected in the embryo, the 2.2-kb transcript was. In addition, the 6Lc hyperploid had a negative effect upon cox3 and atpA in the embryo and cox1 in the endosperm. It would be interesting to determine if these effects were the coincidental result of multiple factors or were due to a single factor with a general effect upon mitochondrial mRNA levels.

A related question concerns the possibility of coordinate regulation of genes that encode functionally related proteins. In yeast the cytochrome c oxidase complex utilizes the various subunits in an equimolar stoichiometric ratio (CAPALDI 1990 ; DROSTE et al. 1996 ). Transcriptional regulation of these genes is coordinated at least in part by the modulation of the level of the nuclearly encoded mitochondrial RNA polymerase (ULERY et al. 1994 ). Our data offered little evidence that dosage-sensitive factors were involved in the coordinate regulation of the RNA levels in maize. The functionally related cox1, cox2, and cox3 genes were no more likely to have coincidental effects with each other than they were with atpA. A possible exception was the effect of the 5Sc hypoploid. It had a negative effect upon all three cox genes in the embryo. Coincidentally, it also had a negative effect upon cox1 and cox2 in the endosperm. The other mitochondrial genes did not share these effects, nor did nuclearly encoded cox5b.

A well-characterized dosage phenomenon relating to the kernel concerns endosperm size. Six of the B-A chromosomes used to create the dosage series carry endosperm size factors (Ef; BIRCHLER 1993 ). Endosperms that are hypoploid for 1Sb, 1La, 4Sa, 5Sc, 7Lb, and 10L19 are markedly smaller than euploid endosperm. The data were examined as to whether effects upon mitochondrial genes could be responsible for the reduced size of hypoploid endosperms. The effects detected in this study did not correlate with Efs. Although this does not rule out the possibility that one or more Efs may result in reduced kernel size because of a limitation of mitochondrial function or a general reduction in mitochondrial gene expression, it seems unlikely that Efs are primarily due to a modulation of mitochondrial transcript level.

Nevertheless, it is possible that a limitation of mitochondrial function contributes to aneuploid syndromes. Indeed, in plants heteroplasmic for mutant mitochondria, the proportion of functional mitochondria correlates with general plant vigor (MARIENFELD and NEWTON 1994 ). Because there are decreases of gene expression in both hypoploids and hyperploids, it is conceivable that these reductions become rate limiting on growth. This can explain the observation that aneuploids have less vigor than do their euploid siblings (BIRCHLER and NEWTON 1981 ; GUO and BIRCHLER 1994 ).

The multitude of effects most likely reflects a hierarchy of dosage-dependent processes. The molecular identification of individual genes that are responsible for such dosage effects indicates a heterogeneous collection of regulatory genes as the responsible agents (e.g., HENIKOFF 1996 ; FROLOV et al. 1998 ). It appears, therefore, that many regulatory mechanisms have evolved to operate in such a manner and at such a level as to be rate limiting on the phenotype. Genes at the top of the hierarchy will still affect the phenotype because the dosage effect will be transmitted through a series of dosage-dependent steps to the ultimate monitored phenotype. This concept unifies the polygenic nature of additive quantitative traits and the extensive effects of aneuploidy on any one characteristic. The large number of modulations of mitochondrial gene expression found in this study is consistent with this view. Also, this type of hierarchy explains a similar spectrum of effects on mitochondrial and nuclear genes despite a fundamentally different chromatin configuration; the hierarchy potentially operates through a series of steps involving nuclear genes before the ultimate impact on mitochondrial expression.

In summary, numerous chromosomal dosage effects were found that resulted in the modulation of the RNA levels of mitochondrial genes. The array of dosage effects upon mitochondrial genes was similar to what had been previously found with nuclear genes. For the most part modulations were tissue specific. A dosage series that affected transcript levels of a gene in embryo tissue usually had a different effect in the endosperm. Most effects were negative; i.e., increased chromosome copy number resulted in lower gene expression levels. This similarity in the array of effects indicates that although mitochondrial genes are prokaryotic in origin, their regulation has been extensively integrated into the eukaryotic cell. Even so, the modes of regulation for mitochondrial and nuclear genes are not identical. A notable difference was that nuclear genes experienced mostly positive effects in the endosperm while the effects upon mitochondrial genes were mostly negative. There is little evidence of coordinate regulation of the cox genes at the RNA level.

	ACKNOWLEDGMENTS

We thank E. Kuzmin, C. J. Leaver, R. M. Mulligan, R. Hiesel, R. Phillips, and Pioneer Hi-Bred Intl. for sharing clones used in this study. We thank M. Guo and K. Yamato for subcloning several of the clones into expressible vectors. We also thank M. Guo for her critical review of this manuscript and C. Ostlie for her grammatical review. D.L.A. acknowledges support from U.S. Department of Agriculture Cooperative State Research, Education and Extension Service grant 96-35304-3492 and the University of Missouri Maize Biology Training Program (a DOE/NSF/USDA Collaborative Research in Plant Biology Program). Research was supported by a grant from the Department of Energy Biosciences Program (to J.A.B.) and the National Science Foundation (to K.J.N.).

Manuscript received October 6, 2000; Accepted for publication December 21, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLEN, J. O., G. K. EMENHISER, and J. L. KERMICLE, 1989  Miniature kernels and plant: interaction between teosinte cytoplasmic genomes and maize nuclear genomes. Maydica 34:277-290.

BECKETT, J. B., 1978  B-A translocations in maize. J. Hered. 69:27-36[Abstract/Free Full Text].

BIRCHLER, J. A., 1981  The genetic basis of dosage compensation of alcohol dehydrogenase-1 in maize. Genetics 97:625-637[Abstract/Free Full Text].

BIRCHLER, J. A., 1993  Dosage analysis of maize endosperm development. Annu. Rev. Genet. 27:181-204[Medline].

BIRCHLER, J. A. and M. R. ALFENITO, 1993  Marker systems for B-A translocations in maize. J. Hered. 84:135-138[Abstract/Free Full Text].

BIRCHLER, J. A. and J. C. HIEBERT, 1989  Interaction of the Enhancer of white-apricot with transposable element alleles at the white locus in Drosophila melanogaster. Genetics 122:129-138[Abstract/Free Full Text].

BIRCHLER, J. A. and K. J. NEWTON, 1981  Modulation of protein levels in chromosomal dosage series of maize: the biochemical basis of aneuploid syndromes. Genetics 99:247-266[Abstract/Free Full Text].

BRAUN, C. J. and C. S. LEVINGS, 1985  Nucleotide sequence of the F1-ATPase alpha subunit gene from maize mitochondria. Plant Physiol. 79:571-577[Abstract/Free Full Text].

CAPALDI, R. A., 1990  Structure and function of cytochrome c oxidase. Annu. Rev. Biochem. 59:569-596[Medline].

CLOWES, F. A. L. and B. E. JUNIPER, 1964  The fine structure of the quiescent center and neighboring tissues in root meristem. J. Exp. Bot. 15:622-630[Abstract/Free Full Text].

CONE, K. C., F. A. BURR, and B. BURR, 1986  Molecular analysis of the maize anthocyanin regulatory locus C1. Proc. Natl. Acad. Sci. USA 83:9631-9635[Abstract/Free Full Text].

COOPER, P., E. BUTLER, and K. NEWTON, 1990  Identification of a maize nuclear gene which influences the size and number of cox2 transcripts in mitochondria of perennial teosintes. Genetics 126:461-467[Abstract].

DROSTE, M., S. MOLLNER, and T. PFEUFFER, 1996  Localisation of an ATP-binding site on adenylyl cyclase type I after chemical and enzymatic fragmentation. FEBS Lett. 391:209-214[Medline].

FAURON, C., M. CASPER, Y. GAO, and B. MOORE, 1995  The maize mitochondrial genome: dynamic, yet functional. Trends Genet. 11:228-235[Medline].

FOX, T. D. and C. J. LEAVER, 1981  The Zea mays mitochondrial gene coding cytochrome oxidase subunit II has an intervening sequence and does not contain TGA codons. Cell 26:315-323[Medline].

FROLOV, M. V., E. V. BENEVOLENSKAYA, and J. A. BIRCHLER, 1998  Regena (Rga), a Drosophila homolog of the global negative transcriptional regulator CDC36 (NOT2) from yeast, modifies gene expression and suppresses position-effect variegation. Genetics 148:317-329[Abstract/Free Full Text].

GRACEN, V. and C. GROGAN, 1974  Diversity and suitability for hybrid production of different sources of cytoplasmic male sterility in maize. Agron. J. 66:654-657[Abstract/Free Full Text].

GRAY, M. W. and B. F. LANG, 1998  Transcription in chloroplasts and mitochondria: a tale of two polymerases. Trends Microbiol. 6:1-3[Medline].

GUO, M. and J. A. BIRCHLER, 1994  Trans-acting dosage effects on the expression of model gene systems in maize aneuploids. Science 266:1999-2002[Abstract/Free Full Text].

HANSON, M. R. and O. FOLKERTS, 1992  Structure and function of the higher plant mitochondrial genome. Int. Rev. Cytol. 141:129-172.

HENIKOFF, S., 1996  Dosage dependent modification of position-effect variegation in Drosophila. BioEssays 18:401-409[Medline].

HERRIN, D. L. and G. W. SCHMIDT, 1988  Rapid, reversible staining of northern blots prior to hybridization. Biotechniques 6:196-197[Medline].

HIESEL, R., W. SCHOBEL, W. SCHUSTER, and A. BRENNICKE, 1987  The cytochrome oxidase subunit I and subunit III genes in Oenothera mitochondria are transcribed from identical promoter sequences. EMBO J. 6:29-34[Medline].

ISAAC, P. G., V. P. JONES, and C. J. LEAVER, 1985a  The maize cytochrome c oxidase subunit I gene: sequence, expression and rearrangement in cytoplasmic male sterile plants. EMBO J. 4:1617-1623[Medline].

ISAAC, P. G., A. BRENNICKE, S. M. DUNBAR, and C. J. LEAVER, 1985b  The mitochondrial genome of fertile maize (Zea mays L.) contains two copies of the gene encoding the alpha-subunit of the F1-ATPase. Curr. Genet. 10:321-328[Medline].

JUNIPER, B. E. and F. A. L. CLOWES, 1965  Cytoplasmic organelles and cell growth in root caps. Nature 208:864-865.

KADOWAKI, K., N. KUBO, K. OZAWA, and A. HIRAI, 1996  Targeting presequence acquisition after mitochondrial gene transfer to the nucleus occurs by duplication of existing targeting signals. EMBO J. 15:6652-6661[Medline].

LAUGHNAN, J. and S. GABAY-LAUGHNAN, 1983  Cytoplasmic male sterility in maize. Annu. Rev. Genet. 17:27-48[Medline].

LEE, S. L. J. and H. E. WARMKE, 1979  Organelle size and number in fertile and T-cytoplasmic male-sterile corn. Am. J. Bot. 60:141-148.

LEVINGS, C. S., III, 1993  Thoughts on cytoplasmic male sterility in cms-T maize. Plant Cell 5:1285-1290[Free Full Text].

LI, X.-Q., M. ZHANG, and G. G. BROWN, 1996  Cell-specific expression of mitochondrial transcripts in maize seedlings. Plant Cell 8:1961-1975[Abstract].

LUPOLD, D. S., A. G. F. S. CAOILE, and D. B. STERN, 1999a  The maize mitochondrial cox2 gene has five promoters in two genomic regions, including a complex promoter consisting of seven overlapping units. J. Biol. Chem. 274:3897-3903[Abstract/Free Full Text].

LUPOLD, D. S., A. G. F. S. CAOILE, and D. B. STERN, 1999b  Genomic context influences the activity of maize mitochondrial cox2 promoters. Proc. Natl. Acad. Sci. USA 96:11670-11675[Abstract/Free Full Text].

MAGUIRE, M. P., 1985  Effective homologous pairing behavior of a tetrasomic chromosome segment in an unusual B-A rearrangement constitution in maize. Caryologia 38:121-128.

MARIENFELD, J. R. and K. J. NEWTON, 1994  The maize NCS2 abnormal growth mutant has a chimeric nad4-nad7 mitochondrial gene and is associated with reduced complex I function. Genetics 138:855-863[Abstract].

MESSING, J., J. CARLSON, G. HAGEN, I. RUBENSTEIN, and A. OLESON, 1984  Cloning and sequencing of the ribosomal RNA genes in maize: the 17S region. DNA 3:31-40[Medline].

MONÉGER, F., C. J. SMART, and C. J. LEAVER, 1994  Nuclear restoration of cytoplasmic male sterility in sunflower is associated with the tissue-specific regulation of a novel mitochondrial gene. EMBO J. 13:8-17[Medline].

MULLIGAN, R. M., A. P. MALONEY, and V. WALBOT, 1988  RNA processing and multiple transcription initiation sites result in transcript size heterogeneity in maize mitochondria. Mol. Gen. Genet. 211:373-380[Medline].

NEUFFER, M. G., 1994 Growing maize for genetic studies, pp. 197–209 in The Maize Handbook, edited by M. FREELING and V. WALBOT. Springer-Verlag, New York.

NEWTON, K. J. and E. H. COE, JR., 1989  Multiple fertility restorers for EP (Z. p.) cytoplasm. Maize Genet. Coop. Newsl. 63:68-69.

NEWTON, K. J., B. WINBERG, K. YAMATO, S. LUPOLD, and D. B. STERN, 1995  Evidence for a novel mitochondrial promoter preceding the cox2 gene of perennial teosintes. EMBO J. 14:585-593[Medline].

ROMAN, H., 1947  Mitotic nondisjunction in the case of interchanges involving the B-type chromosome in maize. Genetics 32:391-409[Free Full Text].

SCHNABLE, P. S. and R. P. WISE, 1998  The molecular basis of cytoplasmic male sterility and fertility restoration. Trends Plant Sci. 3:175-180.

SINGH, M. and G. G. BROWN, 1993  Characterization of expression of a mitochondrial gene region associated with the Brassica "Polima" CMS: developmental influences. Curr. Genet. 24:316-322[Medline].

SMART, C., F. MONÉGER, and C. J. LEAVER, 1994  Cell-specific regulation of gene expression in mitochondria during anther development in sunflower. Plant Cell 6:811-825[Abstract].

ULERY, T. L., S. H. JANG, and J. A. JAEHNING, 1994  Glucose repression of yeast mitochondrial transcription: kinetics of derepression and role of nuclear genes. Mol. Cell. Biol. 14:1160-1170[Abstract/Free Full Text].

YOUNG, E. G. and M. R. HANSON, 1987  A fused mitochondrial gene associated with cytoplasmic male sterility is developmentally regulated. Cell 50:41-49[Medline].

This article has been cited by other articles:

				W. F. Sheridan and D. L. Auger Construction and Uses of New Compound B-A-A Maize Chromosome Translocations Genetics, December 1, 2006; 174(4): 1755 - 1765. [Abstract] [Full Text] [PDF]

				D. L. Auger, A. D. Gray, T. S. Ream, A. Kato, E. H. Coe Jr., and J. A. Birchler Nonadditive Gene Expression in Diploid and Triploid Hybrids of Maize Genetics, January 1, 2005; 169(1): 389 - 397. [Abstract] [Full Text] [PDF]

				J. A. Birchler, D. L. Auger, and N. C. Riddle In Search of the Molecular Basis of Heterosis PLANT CELL, October 1, 2003; 15(10): 2236 - 2239. [Full Text] [PDF]

				M. A. R. Milla, E. Butler, A. R. Huete, C. F. Wilson, O. Anderson, and J. P. Gustafson Expressed Sequence Tag-Based Gene Expression Analysis under Aluminum Stress in Rye Plant Physiology, December 1, 2002; 130(4): 1706 - 1716. [Abstract] [Full Text] [PDF]

				D. L. Auger and J. A. Birchler Maize Tertiary Trisomic Stocks Derived From B-A Translocations J. Hered., January 1, 2002; 93(1): 42 - 47. [Abstract] [Full Text] [PDF]