The Gene Balance Hypothesis: From Classical Genetics to Modern Genomics

The concept of genetic balance traces back to the early days of genetics. Additions or subtractions of single chromosomes to the karyotype (aneuploidy) produced greater impacts on the phenotype than whole-genome changes (ploidy). Studies on changes in gene expression in aneuploid and ploidy series revealed a parallel relationship leading to the concept that regulatory genes exhibited a stoichiometric balance, which if upset, would modulate target gene expression. The responsible regulatory genes for these types of effects primarily have been found to be members of signal transduction pathways or transcription factors of various types. Recent studies of retention of selected duplicate genes following diploidization of ancient polyplodization events have found that signal transduction and transcription factors have been preferentially maintained in a dosage-sensitive relationship. In this essay, we review the historical progression of ideas about genetic balance and discuss some challenges in this field for the future. Chromosomal Imbalance. Diagrammatic representation of chromosome variations with the varied chromosome shown in enlarged form. Full or partial monosomics with only one copy of a chromosome or chromosomal segment are usually more severely affected than haploid plants. Trisomic plants with an extra chromosome often show a greater morphological change from the diploid than observed in polyploid plants, such as tetraploids. Secondary trisomics carrying an isochromosome composed of two copies of a chromosome arm show enhanced phenotypic changes for some characteristics typical of the corresponding primary trisomic. When an isochromosome is recovered in an otherwise haploid plant, the imbalance is most severe and results in highly defective plants (Satina et al., 1937). These relationships led to the concept of genetic balance. As this work continued, trisomics were found for all of the possible 12 chromosomes of Datura, but additional forms continued to arise, each with different characteristics. Some of these new forms accentuated a portion of the phenotypic characteristics of the original set of primary trisomics but were missing other changes from the normal diploid. The basis of these forms was revealed to be an extra chromosome that possessed two identical chromosome arms derived from misdivision of the centromere, manifested by a break in the middle of a centromere with fusion of two sister chromatids, to produce mirror image chromosomes (isochromosomes) generated from a single chromosome arm (Belling and Blakeslee, 1924). In these cases, the extra chromosome, called a secondary trisome, would result in a total of four copies of the respective arm, thus providing an explanation for the accentuated partial phenotype compared with the respective trisomic from which it was derived (Figure 1). These results bolstered the concept that a balance of genes was important for the normal phenotype. These and other aneuploid types were summarized in an article entitled “New Jimson Weeds from Old Chromosomes” (Blakeslee, 1934), which cryptically conveys the message that the dosage of chromosomes is important in this case as opposed to genetic variation. Indeed, the more extreme the altered relative dosage of chromosomes, the greater the phenotypic change. This concept is aptly illustrated by the recovery of selected isochromosomes in otherwise haploid plants that had an extreme imbalance and severe phenotype (Satina et al., 1937) (Figure 1). At about the same time as the early studies of Datura were conducted by Blakeslee, Richard Goldschmidt was studying the basis of sex determination in the gypsy moth (Lymantria dispar) (Goldschmidt, 1920). Crosses between different geographic isolates produced progeny that had differing levels of intersexuality, which is a mixture of male and female morphologies. To explain the various results, Goldschmidt hypothesized different male and female factors of varying strengths whose balance was critical for proper sex determination. Within a particular geographic isolate, the sex determination processes are normal, but crosses between different strains produced the intersexual individuals because the various factors had varied in dosage potential between the diverged populations and were no longer matched. In retrospect, this result foreshadowed the concept of Muller-Dobzhansky coevolving gene complexes that interact within a species but cause sterility or inviability in hybrids. In further parallel studies, Calvin Bridges identified a triploid female Drosophila and among her progeny found intersexual flies that possessed a mixture of male and female parts (Bridges, 1925). The karyotype of these intersexual flies was composed of two X chromosomes with an otherwise triploid complement of chromosomes. However, flies with only one X chromosome that were otherwise triploid were male-like in phenotype. Of course, among diploid flies, the males have one X chromosome and the females have two, with all other chromosomes being present in duplicate in the two sexes. By comparing the various chromosomal constitutions of males, intersexes, and diploid and triploid females, Bridges concluded that a balance between the X and the autosomes (non-sex chromosomes) was the chromosomal basis of sex determination in Drosophila. This type of mechanism for sex determination has been postulated for many plant species as well, with the most well-documented case being Rumex acetosa (Ono, 1935). When the X-to-A ratio is 1, a female is produced, but when it is 0.5, a male occurs. In polyploid Rumex, X-to-A intermediate ratios produce intersexes and hermaphrodites (Dellaporta and Calderon-Urrea, 1993). Similar results of chromosome additions to produce trisomics and tetrasomics as well as the reciprocal production of monosomic individuals have recapitulated the balance conclusions in scores if not hundreds of species in both the plant and animal kingdoms (e.g., Lee et al., 1996; see discussion in Birchler et al., 2001). Although great variation occurs, in general, both monosomics and trisomics cause detrimental effects on the phenotype, with the monosomics usually being more severe. Aneuploid studies in polyploids have further supported the balance idea because there is less effect of extra or missing chromosomes when the remainder of the genome is increased in copy number. The extensive studies of aneuploids in hexaploid bread wheat by Ernie Sears exemplify these conclusions (Sears, 1953, 1954). Over the subsequent decades, the interpretation of genetic balance split into an interesting dichotomy of enzymatic/metabolic versus gene regulatory balance that lingers today. On the one hand, the involvement with sex determination mechanisms was viewed as a regulatory balance, while on the other hand, aneuploid syndromes were viewed as perturbations of metabolism resulting from altered relationships of enzymes involved with intermediary metabolism. Indeed, the basis of sex determination in Drosophila involves a dosage relationship between transcription factors on the X chromosome relative to those on the autosomes (Erickson and Cline, 1993). While aneuploid syndromes no doubt have a complicated basis, the emerging evidence suggests that it is likely that most of these balance phenomena reflect gene regulatory mechanisms in some form, although other gene products involved with macromolecular complexes are likely to exhibit a balance as well, whether or not they are directly involved with gene expression (see below). Several decades ago, one of us (J.A.B.) examined the levels of enzyme activities and proteins in aneuploid series of maize. A popular exercise at the time was the localization of the cytological position of genes encoding various enzymes by screening the genome using a set of trisomics or segmental trisomics created by overlapping translocations (e.g., Carlson, 1972; O’Brien and Gethman, 1973). Work on Datura played an inspiring role in this trend (Carlson, 1972). The principle involved the concept that varying the dosage of a gene would produce a directly proportional amount of gene product (Grell, 1962). Thus, reversing the procedure in a screen of trisomics covering the genome should locate the position of the structural loci for various enzymes because the increased dosage of the encoding gene would produce a greater amount of gene product. This approach was successful in some cases; the unsuccessful attempts remain unknown. Thus, to test ideas about the regulation of the Alcohol dehydrogenase-1 gene in maize (Schwartz, 1971), a dosage series of the chromosome arm on which this gene is located was examined for the levels of ADH activity (Birchler, 1979). Interestingly, the total amount of ADH present in a one to four dosage series was nearly equivalent to the diploid level. In other words, no gene dosage effect for Adh was found in the whole arm dosage series. However, a whole genome ploidy series involving monoploids, diploids, and tetraploids indicated that there was a directly proportional amount of ADH per cell through the series. When other enzymes encoded on other chromosomes were examined in the aneuploid series, their levels were modulated, either up or down, but the most common effect was an inverse correlation between the chromosomal dosage and the amount of activity present (Birchler, 1979). These studies were extended to other proteins and other aneuploid and ploidy series with similar results (Birchler and Newton, 1981). A hypothesis was formulated that the stoichiometry of regulatory genes was influential in modulating the levels of expression of the target genes studied (Birchler and Newton, 1981). The failure to find a dosage effect for ADH was referred to as dosage compensation. The basis of this response was determined to be the result of a structural gene dosage effect of Adh itself being modulated in an inverse manner by a different part of the same chromosome arm involved in the aneuploid series (Birchler, 1981). In monosomics, the single Adh gene was upregulated approximately twofold. In trisomics, the three copies of Adh were each downregulated by two-thirds. Of course, dosage compensation of the X chromosome in males of Drosophila had been documented since the realization of its existence by Muller (1932). That is, males with only one X chromosome produce about the same amount of gene product as do the two X chromosomes present in females. However, Devlin et al. (1982) found that compensation would also occur for large trisomics of autosomal chromosome arms consisting of ∼20% of the whole genome, in contrast with small trisomics surrounding a particular structural gene that vary only a very small fraction of the genome, as noted above. Later, they reported the extensive presence of the inverse dosage effect of these large trisomics on genes encoded on other chromosomes (Devlin et al., 1988). The basis of autosomal dosage compensation was also found to involve the combination of a structural gene dosage effect and an inverse effect being produced simultaneously by a large trisomic segment surrounding the Adh gene in flies. In other words, a structural gene dosage effect is cancelled via an inverse effect produced by another part of the varied region (Birchler et al., 1990). In maize, a survey of aneuploids from many regions of the genome for six genes revealed that the RNA levels of any one gene could be modulated similarly by different dosage series (Guo and Birchler, 1994). In other words, changes in dosage of several different segments of the genome could have the same effect on the monitored gene. The magnitude of these modulations depended on whether the tissue examined was diploid or triploid. Adding or subtracting a chromosome arm to the genotype at different ploidy levels produced an effect whose severity was coincident with the degree of dosage imbalance. Changes in gene expression in a whole genome ploidy series were not as prevalent (Guo et al., 1996). Transcriptome studies of mammalian trisomics or highly aneuploid cancer cells indicate extensive transacting modulations of gene expression (e.g., Phillips et al., 2001; FitzPatrick et al., 2002; Saran et al., 2003; Tsafrir et al., 2006). Thus, the types and mode of effects parallel the classical phenotypic studies with regard to balance. A simple interpretation of the cause of these transacting dosage effects is that they are caused by a gene or genes on the varied chromosome that exhibit a dosage effect themselves and that act in a regulatory fashion to modulate many targets. To screen for single gene mutations that would mimic the dosage effect, mutageneses were conducted in Drosophila to find mutations that could up- or downregulate the expression of the white eye color gene using a leaky phenotypic reporter called white-apricot. The eyes of these flies have a low level of pigment, so modulations of expression could be scored easily. An amazingly large number of mutations were recovered from these screens, totaling 47 at last count (Birchler et al., 2001). The first reported example was Inverse regulator-a (Rabinow et al., 1991). Mutations of this gene as a heterozygote upregulate the white gene twofold, mimicking how a monosomic would produce a similar aneuploid inverse of the region downregulate The presence of the in an otherwise triploid white to of the Thus, as an example of a single gene that a balance effect, in this case with an inverse dosage The large number of of a single target gene be in the that often in a That is, one early a and so each is dosage the effect could be through the also that any one many so there would be of for different A of the of the various of white revealed that their basis into two members of signal transduction pathways and factors (Birchler et al., 2001). Thus, it was that the dosage effects were in the result of regulatory However, the of regulatory involved was at the and to some is, but for this is the case are emerging from studies of in and (see below). to the phenotypic effects of trisomics, it is noted that any one of an be affected by different This realization suggests that dosage-sensitive genes be of modulating a particular phenotypic the field of but we discuss some as noted (Guo and Birchler, Birchler et al., any one there often be loci that its expression 1993). Crosses between that for such usually exhibit an intermediate phenotype to some degree 1993). Thus, the loci act as if they are dosage An example involves the genetic between and et al., loci were identified that were of small effect and in mode of that the effects of aneuploids and of some loci have a basis in would be to be a but it is that variation in regulatory genes of types would be to be a Within the it that many result not from a or a but from of particular gene In other words, mutations as a heterozygote would syndromes in a type of dosage As the basis of these was they primarily of transcription To explain these results, one of us formulated of transcription into the of stoichiometry of the for the of the whole This concept has been extended to genetic and The balance between either the of a or between proteins with such as and to be maintained to some to evidence of this type of relationship has not been in plants. However, have been from the of in et al., and and gene were examined in diploid for that produced a were among of genes whose products are involved with While these of genes those involved with regulatory processes to some regulatory gene products this The found that members of signal transduction and were among factors The concept of dosage balance a relationship between the number of of a and to the of dosage effects when or et al. have shown that the number of a has within a genetic variation of gene expression in and they found a correlation between the variation of gene expression and the number of As the of variation in expression among genes encoding was than that of of the existence of a balance relationship. the expression levels of for proteins in the same has been reported several et al., 2002; et al., These results the existence of a dosage balance, which could be a gene expression at a small that is, within in complexes likely in other approximately a et al., which that most are Indeed, of genes involved in complexes only one with the of the such as A or in By only of have or more of from et al., Mutations in the are more likely to to less with of these are to be less dosage-sensitive they are several per as A in This concept explain of of complexes is usually well This is from the of et al. in a of each an copy of a different gene. that in reflect regulatory they found that of genes the cell is more likely to cause cell or than cell This is so because in many these factors in involving a and a on a common effects in these are by dosage balance. using a of the cell et al. found that of the be at in either or However, the of the do not exhibit this and some are very to dosage for the of the are otherwise there be and and the cell a The for the of The for balance between number or of and its copy number or expression of a be to cell that these within the not produce cell but either or As noted the same dosage effects on target gene expression that result from imbalance are responsible for dosage compensation of various genes in maize (Birchler, Birchler and Newton, and Birchler, 1994). The magnitude of expression to for dosage compensation of the various sex chromosome aneuploids in Drosophila is the inverse ratio of the X to imbalance (Birchler et al., 2006). Recent studies of gene expression between males and females of and that in each species the single X chromosome in males is upregulated approximately et al., which would produce a total gene expression equivalent to the two X chromosomes in females hermaphrodites in In an of gene expression of X chromosomes and autosomes in various mammalian species and and a potential balance relationship was In female one of the two X chromosomes is in any one the X in females is upregulated approximately so that on the total expression from the single X is more or less equivalent to the expression of the equivalent of a of The single X in males is However, in haploid the expression of the single X chromosome per is the same as a single Thus, it that the X only when an imbalance is chromosome dosage compensation has been to and thus is likely to involve some in mechanism to those observed in aneuploids (Birchler et al., but some do to the of whole-genome it has that of by diploidization have in the leading to the of present 2001; et al., 2002; et al., the production of an from two there is gene that back to the diploid level. An of the of the genes in duplicate in and an of members of signal transduction and transcription factors and et al., et al., and et al., 2006). These of genes are similar to those that exhibit transacting dosage as noted above. These led to the hypothesis that the are because they are in balance with each other and of one of a would their (Birchler et al., and 2006). In other words, of one of a duplicate would mimic an aneuploid effect, which would not in an balance relationship would be at time back to the diploid level. Indeed, gene preferentially found in segmental are the complement of those from whole-genome and et al., that segmental that a regulatory balance would be selected because they also would mimic an aneuploid Of course, genes in segmental could be selected to greater of a particular gene product et al., et al., 2006). members of large gene of macromolecular complexes in and to studies, of different genes as the of the complexes et al., 2003; et al., The studies on duplicate retention in and by with that are to the diploid Thus, the that of members of the to produce the observed However, a of the genome of the processes of whole-genome and subsequent gene et al., 2006). The of this with a number of that of The for this number is that the whole genome has three By comparing the and into the processes of gene could be many could be in various of it likely that the gene are the result of time than a following the to be of genes that are involved in macromolecular In the genes show evidence of that of one of a is selected because it the stoichiometric relationships of of such The thus that the stoichiometry of is important for their and have that the retention of duplicate regulatory genes following events such in the for time to that The the copy number of all time the copy number of most target genes to the diploid level. However, the stoichiometric on regulatory in the of of different members of a duplicate of and While there is evidence that such occurs, of how a new balance is have been However, such and that the of and diploidization are a to the of A regulatory in which the are to stoichiometric relationships a for to on new mutations in and tetraploids while such mutations are a of genes any one and to a dosage a to modulate the phenotype in via mutations in the regulatory is dosage any new regulatory mutations have the potential to produce effects as a they are they be selected However, new mutations would be in the for changes as The of on in et al., suggests that this type of is The balance relationship for changes to be of small magnitude and to result from many possible of a single phenotype, as noted for the example above. Thus, a balance relationship of regulatory complexes would and for with diploid and levels of have summarized the historical and evidence a balance relationship of regulatory genes as a of their in macromolecular complexes or for which the members produce a stoichiometric effect on the of the are several that such a hypothesis is to the and of macromolecular complexes in an to the basis of how varying the dosage of the of the This involve the of regulatory the in which they and the between the different studies would test the that highly proteins in complexes to a dosage and that this be modulated by the of the one would to the of gene expression on the phenotype. no doubt be using gene it be possible to between of of of genes and their corresponding phenotypic The between aneuploid effects and are noted but the of this relationship is at studies of of gene expression to be both on aneuploid and on to test such potential Indeed, modulations of by aneuploid syndromes and involve changes in cell as as changes in gene the of the responsible The for in the is to the relationship of total expression to phenotypic relationships be important to a of the role of regulatory balance in how this balance many changes in phenotype in but is for into this is to to regulatory mutations are or to each The of regulatory genes among the on the from to in has an for the of In are among regions of the genome et al., 2006). be of to whether they are following events and whether they are to the balance characteristics of regulatory the stoichiometry of regulatory factors to and the be to from a balance with other factors and a new balance. such is a to the processes that from one balance relationship to another be classical studies of species have the presence of coevolving gene complexes see discussion in Birchler et al., regulatory complexes have a relationship to these mechanisms would be an interesting of on this in the Birchler is by a from the and by the of is by the et the and the of