Logo of plntcellLink to Publisher's site
Plant Cell. 2007 May; 19(5): 1458–1472.
PMCID: PMC1913735

The Two AGPase Subunits Evolve at Different Rates in Angiosperms, yet They Are Equally Sensitive to Activity-Altering Amino Acid Changes When Expressed in Bacteria[W]


The rate of protein evolution is generally thought to reflect, at least in part, the proportion of amino acids within the protein that are needed for proper function. In the case of ADP-glucose pyrophosphorylase (AGPase), this premise led to the hypothesis that, because the AGPase small subunit is more conserved compared with the large subunit, a higher proportion of the amino acids of the small subunit are required for enzyme activity compared with the large subunit. Evolutionary analysis indicates that the AGPase small subunit has been subject to more intense purifying selection than the large subunit in the angiosperms. However, random mutagenesis and expression of the maize (Zea mays) endosperm AGPase in bacteria show that the two AGPase subunits are equally predisposed to enzyme activity-altering amino acid changes when expressed in one environment with a single complementary subunit. As an alternative hypothesis, we suggest that the small subunit exhibits more evolutionary constraints in planta than does the large subunit because it is less tissue specific and thus must form functional enzyme complexes with different large subunits. Independent approaches provide data consistent with this alternative hypothesis.


DNA and protein sequence conservation is the foundation for the fields of comparative genomics and molecular evolution. In fact, sequence conservation over evolutionary time has long been assumed to depend on the importance of the gene in question for the fitness of the organism. The overall rate of amino acid substitution of a protein is generally thought to depend on the proportion of amino acids necessary for proper function and the phenotypic impact of mutations (Dickerson, 1971; Zuckerkandl, 1976; Wilson et al., 1977; Kimura, 1983; Nei, 1987; Li, 1997). However, the relationship between fitness effects (revealed by mutant phenotypes) and sequence conservation is complex, and exceptionally conserved proteins encoded by genes with mild mutant phenotypes are well known (Braun et al., 1996).

The nature of the relationship between protein sequence conservation and fitness has attracted considerable interest. The expected negative correlation between protein dispensability (viability of mutants with a knockout in the gene encoding the protein) and sequence conservation was found in some studies (Hirsh and Fraser, 2001; Jordan et al., 2002; Liao et al., 2006), while other studies were either unable to support this correlation or found that it was weak after controlling for other factors (Hurst and Smith, 1999; Rocha and Danchin, 2004). Conservation has also been linked to protein structure (Bloom et al., 2006), protein expression level (Pal et al., 2001; Drummond et al., 2005), tissue specificity (Duret and Mouchiroud, 2000; Akashi, 2001; Wright et al., 2004), gene compactness (Liao et al., 2006), numbers of interacting proteins (Fraser et al., 2002), and position in an interaction network (Hahn and Kern, 2005). There have only recently been efforts to examine the contribution of each of these factors to sequence conservation in a unified framework (e.g., Wall et al., 2005; Wolf et al., 2006).

The heterotetrameric plant enzyme ADP-glucose pyrophosphorylase (AGPase) presents an excellent system in which one can study the functional basis for protein sequence conservation, since it is composed of two identical small and two identical large subunits but the small subunit exhibits strikingly less sequence divergence. Despite the differences in the degree of sequence conservation for the small and large subunits, the subunits are clear paralogs encoded by genes arising from an ancient duplication. The difference in sequence conservation has led others to propose that the two subunits play vastly different roles in enzyme catalysis (Greene et al., 1996; Ballicora et al., 2003, 2005; Frueauf et al., 2003).

AGPase is an allosteric enzyme that catalyzes a rate-limiting step in starch synthesis, the conversion of glucose-1-P and ATP to ADP-glucose and pyrophosphate. ADP-glucose then serves as the main if not sole precursor for starch synthesis by starch synthases (Hannah, 2005). Angiosperm and green algal AGPases are heterotetramers consisting of two identical large and two identical small subunits. In the case of the maize (Zea mays) endosperm AGPase, the subject of this report, the large subunit is encoded by shrunken-2 (Sh2) and the small subunit by brittle-2 (Bt2) (Bae et al., 1990; Bhave et al., 1990). SH2 and BT2 show considerable amino acid identity (43.2%) and similarity (61%) (see Supplemental Figure 1 online). They also exhibit identity to the sole subunit of the homotetrameric Escherichia coli AGPase (SH2, 24.4%, BT2, 29.0%). Gene duplication early in the evolution of the plant lineage followed by sequence divergence is the most parsimonious explanation. The two subunits are not functionally interchangeable, as shown by mutant analysis. Loss of Sh2 or Bt2 function abolishes >90% of endosperm AGPase activity (Hannah and Nelson, 1976). Expression of each of the maize endosperm AGPase subunits separately in E. coli showed that SH2 or BT2 alone gives only 3.5 and 2.5% activity, respectively, of the heterotetramer (Burger et al., 2003).

Small subunit proteins show striking conservation among species, while the large subunits are less conserved (Smith-White and Preiss, 1992; see Results). Smith-White and Preiss (1992) suggested that the small subunit has more selective constraints than does the large subunit. The idea that the small subunit has been subject to greater constraint than the large is supported by the higher percentage of identity between cyanobacterial AGPase and the small subunit (Greene et al., 1996). While the exact role of each subunit remains unclear, Ballicora et al. (2003, 2005), Frueauf et al. (2003), and Greene et al. (1996) working with potato (Solanum tuberosum) tuber AGPase suggested that the two subunits have different roles in AGPase function. The small subunit has been proposed to be catalytic, while the large subunit presumably modifies the response to allosteric regulators. According to this hypothesis, this presumed specificity in roles has placed different evolutionary constraints on the two proteins. The large subunit, then, can accumulate more nondeleterious nonsynonymous mutations for AGPase activity than can the small subunit. However, the roles of each AGPase subunit in enzyme function are not clear, since each subunit has activity when expressed alone in E. coli (Iglesias et al., 1993; Burger et al., 2003), and mutations in either subunit affect both catalytic and allosteric properties (Hannah and Nelson, 1976; Cross et al., 2004, 2005; Hwang et al., 2006, 2007). It should also be noted that the regulatory subunits of the Archaeal type H+-ATPase and vacuolar type H+-ATPase evolved at a slower rate compared with the catalytic subunits (Marin et al., 2001).

Here, we show that the higher degree of sequence divergence in the large subunit can be attributed to increased evolutionary constraints on the small subunit. We performed two independent tests to determine whether the difference in evolutionary rates of the two subunits (BT2 and SH2) reflected different sensitivities of the subunits to activity-altering amino acid changes. Our results indicate that SH2 and BT2 are equally predisposed to activity-altering amino acid changes when expressed in one common environment (E. coli) with a single complementary subunit. We propose, therefore, that the AGPase small subunit has more evolutionary constraints in angiosperms than does the large subunit because it is less tissue specific, less redundant, and must form functional enzyme complexes with different large subunits. In support of this hypothesis, we show that mutant small subunit proteins often have differential effects on enzyme activity when paired with different large subunit proteins.


Alignment of SH2 and BT2 with Homologs

A previous study showed that the AGPase small subunit exhibits greater sequence conservation than does the large subunit (Smith-White and Preiss, 1992). To explore the much larger collection of AGPase sequences and update the comparison of the two AGPase subunits, we aligned all available angiosperm AGPase large and small subunits (see Supplemental Table 1 online). The hypervariable N termini of the two subunits were excluded from the alignments (Figure 1). These alignments showed that the average pairwise identity is 70.8% for large subunits and 91.3% for small subunits. More strikingly, 21.6% of SH2 residues and 64.1% of BT2 residues are invariant within the respective family (Figure 1). It is also apparent that the large subunit is more diverse than is the small subunit throughout their sequences (Figure 1).

Figure 1.
Amino Acid Conservation Patterns of BT2 and SH2.

Evolutionary Constraints of the Large and Small Subunits of AGPase

To determine whether the difference in amino acid identity is due to evolutionary constraints or to the timing and pattern of gene duplication events, the overall ratio of nonsynonymous substitutions per nonsynonymous site to synonymous substitutions per synonymous site (ω) for the large and the small subunit genes was calculated. To check for changes in evolutionary constraints within a gene tree that includes the large and the small AGPase subunits from angiosperms, we applied two nested likelihood models to the gene tree in Figure 2A. A model that assumes two different ω values, one for the large and one for the small subunits, was favored over a model that assumes one ω for both subunits by a likelihood ratio test (LRT; P < 0.001) (Felsenstein, 1981; Goldman, 1993; Yang et al., 1995; Huelsenbeck and Rannala, 1997; Hileman and Baum, 2003). The overall ω estimates were 0.087 for the large subunit and 0.032 for the small subunit (Figure 2A). This result indicates that the large subunit evolved under reduced purifying selection compared with the small subunit. To monitor for saturation of the rapidly accumulating synonymous substitutions that might cause bias, ω values were calculated for sets of closely related isoforms within each subunit. The small subunit genes were clustered into two major groups (Figure 2B). Groups 1 and 2 consist of genes expressed in monocots and eudicots, respectively. The large subunit genes were clustered into five major groups (Figure 2C). Group 1 comprises genes expressed in monocot and eudicot leaf tissue. Group 2 includes genes expressed in eudicot source and sink tissues. Groups 3 and 4 contain genes expressed in monocot seed and eudicot sink tissues, respectively. Group 5 is composed of only two sequences whose tissue specificity has not been studied in detail. A model that assumes a different ω value for each group of the small subunit was favored over a model that assumes a single ω for all small subunit groups (LRT; P < 0.001). The same is true for the large subunit (LRT; P < 0.001). The ω values of the large subunit groups ranged from 0.073 to 0.132 (Figure 2C), while small subunit groups varied from 0.027 to 0.054 (Figure 2B). These results indicate that mutational saturation is unlikely to have had a major impact on calculated ω values and further suggest that the large subunit has been subjected to fewer constraints than has the small subunit.

Figure 2.
Estimates of AGPase Subunit Gene Trees and ω Values.

Probability That a Nonsynonymous Mutation in Sh2 or Bt2 Affects Enzyme Activity

Two Sh2 and two Bt2 libraries were created by error-prone PCR. The resulting sh2 and bt2 clones were expressed in E. coli strain AC70R1-504 with a wild-type complementary subunit, and 96 colonies from each of the four libraries were chosen at random. Colonies were rated as functional or nonfunctional by formation of brown-staining glycogen following exposure to iodine vapors. We expect that mutations that modify catalytic and regulatory properties, enzyme stability, and enzyme assembly would affect enzyme activity. DNA sequencing revealed the nature and position of nucleotide changes. Mutations resulting in activity loss are shown in Supplemental Table 2 online. Mutants listed are those that come from nonfunctional clones containing only a single missense mutation. Nonfunctional clones with multiple missense mutations were excluded because the causal mutation could not be determined.

The distribution of all nucleotide substitutions within Sh2 or within Bt2 is uniform for all libraries (see Supplemental Figure 2 online). The percentage of conservative missense mutations (see Methods) was 54.8% ± 6.9% and 51.3% ± 6.6% (mean ± 2 × se) in the Bt2 and Sh2 libraries, respectively. These percentages are not significantly different, indicating that the comparison of the robustness of Sh2 and Bt2 to missense mutations is unlikely to be biased by introduction of more conservative changes in either subunit. Clones containing indels, stop codons, or no nonsynonymous mutations were excluded from further analysis.

The probability that a nonsynonymous mutation abolishes gene function was estimated by the formula recently published by Guo et al. (2004). It is termed the X-factor (Xf) throughout the article:

equation M1

where S is the fraction of functional clones, fn is the fraction of clones having n nonsynonymous mutations, and Xf is the probability that a nonsynonymous mutation in Sh2 or Bt2 reduces AGPase activity, leading to no obvious production of glycogen.

Results from the two Sh2 and Bt2 libraries are summarized in Table 1. The Xf for Bt2 is 34.02% ± 0.82% and 33.36% ± 2.27% for Sh2. These values are not statistically different. We conclude that Sh2 and Bt2 show little to no difference in robustness to nonsynonymous mutations with respect to AGPase activity when expressed in E. coli.

Table 1.
Glycogen Production from Randomly Chosen sh2 and bt2 Mutants Expressed in E. coli

In these analyses, we assumed that multiple intragenic nonsynonymous mutations act independently. To test the validity of this assumption, we compared the fractions of clones having one or more nonsynonymous mutations not altering AGPase function to the expected fractions if multiple intragenic mutations act independently (see Supplemental Table 3 online). χ2 goodness-of-fit test indicates that the observed fractions are not significantly different from the expected fractions. This indicates that the assumption is valid for Sh2 and Bt2 for the mutational loads of this experiment.

Glycogen Quantitation

Since iodine staining is largely a qualitative method, the glycogen levels of Sh2 and Bt2 colonies of the second libraries were measured by a phenol reaction. This allowed us to determine the relationship between iodine staining and glycogen content. Glycogen contents were compared with the level produced by wild-type Sh2 and Bt2. Iodine staining is a reliable method for classifying clones for AGPase function (Figure 3). Clones rated as nonfunctional by iodine staining produced <4% wild-type glycogen levels. There was little overlap in the glycogen content values among functional and nonfunctional clones as judged by iodine staining. The small overlap was a result of slight inaccuracies in either iodine staining or glycogen quantification. Only three ambiguous clones, producing ~3 to 5% wild-type glycogen levels, were identified. Inclusion of these clones had little effect on calculated Xf values.

Figure 3.
Glycogen Contents of sh2 and bt2 Mutants Compared with Wild-Type AGPase.

We asked if a relationship exists between the presence of synonymous mutations and glycogen production. Pearson correlation coefficient r is 0.039 for Bt2 and 0.043 for Sh2 (P > 0.05 for both), indicating little to no correlation between the number of synonymous mutations per clone and glycogen production. This result verifies our expectation that synonymous mutations do not affect AGPase activity in this experiment.

Glycogen quantitation also provides the opportunity to change the stringency of classification of functionality. Increasing the stringency of wild-type classification from 0 to 99% glycogen increases Xf values; however, almost identical Xf values for Sh2 and Bt2 are found at each stringency level (Table 2). Hence, regardless of the stringency level imposed, Sh2 and Bt2 are equally susceptible to activity-altering synonymous mutations.

Table 2.
Xf of Sh2 and Bt2 Using Different Stringencies to Classify Mutants as Nonfunctional

Estimating the Ratio of Nonsynonymous to Synonymous Mutations in Functional Clones

As an independent estimate of the robustness of Sh2 and Bt2 to nonsynonymous mutations affecting AGPase activity, we isolated functional clones from heavily mutagenized libraries and measured the ratio of nonsynonymous to synonymous mutations. In our previous studies, above, the ratios of nonsynonymous to synonymous mutations of randomly selected Sh2 (2.936 ± 0.1037) and randomly selected Bt2 (3.0585 ± 0.063) clones are not significantly different. Hence, synonymous mutations in functional clones can be used as a comparison of mutational load between Sh2 and Bt2 in this experiment.

Accordingly, two additional Sh2 and two additional Bt2 heavily mutagenized libraries were created by error-prone PCR. Resulting clones were expressed in E. coli AC70R1-504, and 48 functional clones from each of the four libraries were sequenced. The number of synonymous mutations per 1000 nucleotides of Sh2 and Bt2 cDNA is 0.951 ± 0.107 and 0.989 ± 0.204, respectively (Table 3). This indicates that the average mutational frequency is the same for both genes.

Table 3.
Ratio of Nonsynonymous to Synonymous Mutations in Functional Sh2 and Bt2 Clones

The ratio of nonsynonymous to synonymous mutations is 1.188 ± 0.036 for Sh2 and 1.176 ± 0.006 for Bt2 (Table 3). These are not significantly different. This indicates that Sh2 and Bt2 are equally robust to nonsynonymous mutations with respect to AGPase activity.

Additionally, χ2 goodness-of-fit tests indicate that the distribution of functional Bt2 clones with different numbers (e.g., 0, 1, 2, 3) of nonsynonymous mutations was not significantly different from the distribution of functional clones found in Sh2 (see Supplemental Table 4 online). This result is in agreement with the conclusion that Sh2 and Bt2 do not differ in their ability to tolerate functional nonsynonymous mutations.

To summarize, two independent tests of protein robustness of the SH2 and the BT2 protein were performed. Both tests lead to the conclusion that the genes are equally predisposed to non-/less-functional nonsynonymous mutations at least when expressed in one environment (E. coli) and in the presence of a single functional complementary subunit.

Proteins Interacting with BT2

We tested the hypothesis that relative to the large subunit the small subunit is more conserved because it plays additional roles that are mediated through protein–protein interactions. To investigate this possibility, we searched for proteins interacting with BT2 in maize endosperm. Formaldehyde was used to cross-link associated proteins in maize endosperm cells using methodology described elsewhere (Hall and Struhl, 2002; Rohila et al., 2004). Cross-linking with formaldehyde traps the BT2 protein into a complex of ~250 kD (Figure 4A). Purification of BT2 from maize endosperm by a monoclonal BT2 antibody column shows that SH2 is the major protein that interacts with BT2 (Figure 4B) as verified by protein gel blot analysis using a polyclonal SH2 antibody (data not shown). The extra proteins that are copurified with BT2 either interacted directly with BT2 or AGPase or they were cross-reaction products with the monoclonal BT2 antibody column (Figure 4B). To distinguish among these possibilities, the same monoclonal BT2 antibody column was used to purify cross-linked proteins from the developing endosperm of bt2-7410, a bt2 knockout mutant that has undetectable amounts of BT2 (Giroux and Hannah, 1994). Six of the eight extra proteins purified from the wild type were also purified from bt2-7410 (Figure 4B). This indicates that these extra proteins are a cross-reaction product with the BT2 antibody column rather than BT2 or AGPase interacting proteins. The remaining two extra proteins purified from the wild type, but not bt2-7410, are SH2/BT2 polymers since they hybridize to both SH2 and BT2 antibodies (data not shown). These results show that SH2 is the only protein interacting with BT2 in endosperm. They also suggest that the AGPase small subunit is unlikely to have additional functions mediated by protein–protein interactions.

Figure 4.
Purification of BT2 Using a Monoclonal Antibody Column.

Tissue Specificity and Gene Number as a Cause of the Greater Evolutionary Constraints on AGPase Small Subunit

The timing of gene duplications giving rise to tissue-specific isoforms differed dramatically for the two subunits (Hannah et al., 2001). At least one duplication of the large subunit occurred before the monocot-eudicot separation, while small subunit duplications occurred after this divergence. The original small subunit duplications appear to have occurred independently in monocots and eudicots and after their evolutionary separation.

The number of successful duplication events also appears to differ for the two subunits. Perusal of the rice (Oryza sativa), Populus (Populus trichocarpa), and Arabidopsis thaliana completed genomes suggests that the large subunit genes underwent more successful duplications than did the small subunit genes. The Arabidopsis genome contains a single functional gene for the AGPase small subunit but four genes for the large subunit (Crevillen et al., 2003, 2005), rice contains two small subunit genes and four large subunit genes (Akihiro et al., 2005; Ohdan et al., 2005), and Populus apparently has one small and six large subunit genes (Tuskan et al., 2006). Three large subunit genes and one small subunit gene have been isolated from tomato (Solanum lycopersicum) and potato (La Cognata et al., 1995; Park and Chung, 1998). The major small subunits in barley (Hordeum vulgare) endosperm and leaves (Thorbjørnsen et al., 1996; Johnson et al., 2003; Rosti et al., 2006) are encoded by a single gene that is alternatively spliced, while the interacting large subunits are encoded by different genes. Expression patterns and functional characterization of large and small subunits from tomato, potato, barley, rice, and Arabidopsis indicate that large subunits are more tissue specific than are small subunits (La Cognata et al., 1995; Park and Chung, 1998; Akihiro et al., 2005; Crevillen et al., 2005; Ohdan et al., 2005; Rosti et al., 2006). Taken as a whole, these results strongly suggest that the large subunit genes, in comparison to small subunit genes, underwent more successful duplications that were then followed in some cases by subfunctionalization in their expression pattern.

There is ample evidence that broadly expressed genes are more conserved than are tissue-specific genes in plants and mammals (Duret and Mouchiroud, 2000; Akashi, 2001; Wright et al., 2004). Broadly expressed genes, like the small subunit AGPase genes, may have additional constraints because they have a greater impact on fitness and/or because they must function in multiple cellular and subcellular environments (Akashi, 2001).

As a result of broad expression, the small AGPase subunit must form an enzyme complex with multiple large subunits in multiple cellular environments. This is not the case for the large subunits. If this hypothesis is true, the effect of small subunit nonsynonymous mutations on AGPase activity should depend on the identity of the large subunit. Mutations in the small subunit that are tolerated in a complex with one large subunit may not be tolerated in a complex with a different large subunit.

To test one aspect of this hypothesis, we randomly selected 20 Bt2 variants that function with the SH2 protein and expressed them in E. coli with the wild-type maize embryo large subunit (Agplemzm) (Giroux and Hannah, 1994; Giroux et al., 1995). Resulting glycogen was measured and compared with the glycogen produced by cells expressing wild-type Bt2 with Sh2 and wild-type Bt2 with Agplemzm, respectively. Cells expressing wild-type Bt2 with Sh2 and wild-type Bt2 with Agplemzm produce approximately equal amounts of glycogen (data not shown). Approximately 25% of Bt2 mutants result in significantly different amounts of glycogen depending on the large subunit partner (Figure 5A). Staining of two variants is also shown (Figure 5B). These results indicate that the effect of some amino acid changes in the small subunit on AGP activity depends on the interacting large subunit.

Figure 5.
Comparison of Glycogen Produced by Bt2 Mutants Expressed with Wild-Type Sh2 and Wild-Type Agplemzm in E. coli.

To further test the hypothesis that the small subunit is more constrained because of broad tissue expression, we perused the evolutionary relationships of AGPase subunits from the sequenced unicellular green algae Chlamydomonas reinhardtii, Ostreococcus tauri, and Ostreococcus lucimarinus. There is one major large and one small AGPase subunit in C. reinhardtii (Zabawinski et al., 2001). This is most likely the case in O. tauri (Ral et al., 2004) and O. lucimarinus since a third gene homologous to an AGPase subunit in these genomes shows very low identity (~31%) to the large and small subunits from the respective organisms. Additionally, alignment of the divergent Ostreococcus AGPase homologs to AGPase subunits from angiosperms, algae, and bacteria shows that residues conserved in all AGPase subunits are not conserved in these extra subunits (data not shown).

In support of the hypothesis that the greater conservation reflects the existence of multiple large subunit genes and/or broad patterns of expression, the large and small subunits in these green algae show similar levels of sequence conservation (see Supplemental Table 5 online). Indeed, the small subunit in these algae evolved more rapidly than did the small subunit in plants and at the same rate as the large subunit of the algae (Figure 6). It should be noted that the rate of evolution of nuclear genes in green algae and in angiosperms is comparable based on small and large subunit rRNA (Van de Peer et al., 1996; Ben Ali et al., 2001) and proteins such as the elongation factor 1α, actin, α-tubulin, and β-tubulin (Baldauf et al., 2000). Hence, the most parsimonious explanation for the faster evolution of the small subunit in algae is that it has fewer constraints in algae than it does in angiosperms.

Figure 6.
Phylogenetic Tree of AGPase Isoforms from Angiosperms and Unicellular Green Algae.

It is also possible that constraints on the AGPase subunits depend on the identity of tissue(s) of expression. Interestingly, large subunits expressed primarily in leaves have lower ω values than do large subunits expressed in monocot seed (Figure 2C). Further support for more evolutionary pressure on subsets of AGPase genes comes from analysis of a recent duplication of a small subunit maize gene. While rice, barley, and probably wheat (Triticum aestivum) have one small subunit gene expressed in both leaves and endosperm, maize has one small subunit gene specific for endosperm and one specific for leaves (Fuchs, 1977; Prioul et al., 1994). These two maize isoforms were created by recent gene duplication. A model that assumes a separate ω value for each branch of the gene tree whose topology is shown in Supplemental Figure 3 online is favored over a model that assumes one ω for the whole tree (LRT; P < 0.01). Data presented in Supplemental Figure 3 online suggest that following duplication, the maize leaf isoform (Z. mays 2) evolved slower than did the maize endosperm isoform (Bt2). Also, both isoforms appear to have evolved faster following duplication (see Supplemental Figure 3 online). These data are in accord with the idea that tissue identity influences the rate of evolution of both the large and the small subunits of AGPase. Expression in multiple tissues then would compound the various selection pressures of the various tissues.


Differential Constraints on the Small and Large AGPase Subunits and Robustness to Activity-Altering Amino Acid Changes

We focused on the cause of the different sequence conservation exhibited by the two subunits of angiosperm AGPase. Substantially greater sequence divergence is found among large subunits vis-à-vis small subunits. Here, we show that the difference in sequence divergence between the large and small AGPase subunits is due to different evolutionary constraints rather than different paths of gene evolution.

One proposed explanation for the differences in sequence conservation is the conjecture that the two subunits play different roles in enzyme catalysis. It has been proposed that the small subunit is catalytic and, accordingly, more constrained in amino acid substitutions compared with the principally regulatory large subunit. This model predicts that random nonsynonymous mutations in the small subunit have a greater probability of disrupting AGPase activity compared with random nonsynonymous mutations in the large subunit.

Two independent mutagenesis studies were used to test whether the large subunit is more robust to mutations than is the small subunit in terms of AGPase activity. The first experiments revealed that random, nonsynonymous mutations in either subunit have the same probability of altering AGPase activity. Approximately one-third of missense mutations reduced glycogen content to ~4% wild-type levels. In the second series of experiments, only functional clones were selected and the frequencies of nonsynonymous and synonymous mutations were determined. If random, nonsynonymous mutations were tolerated at a higher frequency in Sh2, then we would expect a higher nonsynonymous/synonymous ratio for Sh2 compared with Bt2. By contrast, we observed the same ratio for both genes.

The conclusion from both sets of experiments is that the probability that a nonsynonymous mutation affects AGPase activity is the same for both the large subunit and the small subunit of AGPase when expressed in one common environment with one complementary partner. This indicates that different roles of the large and small subunits in AGPase activity, if they exist, cannot account for the increased evolutionary constraints placed on the plant small subunit. This is also supported by the fact that in green unicellular algae the large subunit and small subunits evolve at similar rates.

Roles of the Large and Small Subunits in AGPase Activity

The proposition that the small subunit is catalytic and the large subunit is involved in regulation is based on three observations. The first is that mutations of some small subunit residues presumably involved in catalysis had greater effect on activity compared with mutants in cognate sites in the large subunit (Fu et al., 1998; Frueauf et al., 2003). The second observation is that the small subunit had activity as a homotetramer, while the large subunit did not (Ballicora et al., 1995). Finally, the small subunit shows more sequence conservation than the large subunit.

The first observation has been challenged by biochemical studies (Hannah and Nelson, 1976; Cross et al., 2004, 2005; Hwang et al., 2006, 2007). Mutations in the large subunit affect catalytic properties and allosteric ones. With regard to the second observation, it should be noted that activity arising from large subunit homotetramers can be seen in the data of Burger et al. (2003) and Iglesias et al. (1993). While activity arising from a large subunit homotetramer is not universally accepted, the lack of activity of a large subunit homotetramer does not show that the large subunit is not catalytic in a heterotetramer.

The third observation, that the small subunit is more evolutionarily conserved than is the large subunit because the small subunit participates in catalysis whereas the large subunit simply modulates regulation, predicts that we should have found more missense mutations affecting activity in the small subunit (compared with the large subunit). By contrast, we found that both proteins were equally predisposed to activity-altering missense mutations. Taking into account the biochemical studies showing that both subunits are important for catalytic and allosteric properties, we believe the simplistic separation of subunit roles to catalytic and allosteric modules is no longer tenable.

Alternative Hypotheses Explaining the Shift in Selective Constraints on the Large and the Small Subunits in Angiosperms

Because both subunits exhibit equal constraints on amino acid sequence when both subunits are expressed in one common environment with a complementary subunit, other explanations must account for the differences in amino acid substitution rates. We considered the possibility that the small subunit plays biological roles separate from its function as an AGPase subunit. Here, we show that SH2 was the only protein that interacted with the BT2 protein in the maize endosperm. Thus, there is no evidence showing that BT2 has functions separate from interacting with SH2 to form an active AGPase.

Several lines of evidence reviewed above suggest that there are fewer copies of small subunit genes than large subunit genes and that small subunit genes have a broader expression profile compared with large subunit genes. Accordingly, evidence that broadly expressed genes evolve at a slower rate than do tissue-specific genes (Duret and Mouchiroud, 2000; Akashi, 2001; Wright et al., 2004) is relevant. Investigations of various gene families showed that genes encoding broadly expressed proteins evolve more slowly than genes with narrow expression profiles (Hastings, 1996; Schmidt et al., 1997). This may reflect a larger fitness effect of mutations that impair gene function in many different tissues or it may reflect more constraints on proteins expressed in multiple biochemical environments (Akashi, 2001). Our studies are consistent with the small subunit having more constraints because it functions in multiple cellular environments and polymerizes with different large subunits.

Although the small subunit does not have interaction partners other than the large subunit, the existence of multiple large subunits that interact with a single small subunit makes the correlation between numbers of interaction partners and conservation relevant (e.g., Fraser et al., 2002). The relationship between conservation and numbers of interaction partners is complex (Jordan et al., 2003), making it difficult to predict a priori whether interaction with multiple paralogous large subunits provides additional constraints. However, when we asked whether BT2 variants that function with the SH2 protein also function with another wild-type large subunit isoform, we found that 25% of randomly selected BT2 variants interact differently with the second large subunit. This provides empirical evidence that large subunit–small subunit interactions differ based on the large subunit involved and indicates that the E. coli experiments did not consider all mutations that are relevant in planta. This reflects the fact that mutations that are deleterious in only one of the various complexes could reduce the fitness of the organism and be eliminated by natural selection.

The hypothesis that tissue specificity and the need to interact with multiple large subunits constrains evolution of the small subunit is also supported by the fact that the small subunit evolved as rapidly as the large subunit in unicellular algae C. reinhardtii, O. tauri, and O. lucimarinus and that the evolution of the small subunit was considerably faster in these algae than it was in angiosperms. Since there is a single cell type in these algae (although it can undergo developmental changes) and there is likely to be a single large subunit, one would predict that the evolutionary rates for the large and small subunits would be identical in these organisms. This is precisely what was observed. If different constraints in evolution of the small and the large subunits were due to different roles in AGPase catalysis, we would expect to see different conservation between the small and the large subunit in algae. The fact that the small subunit in unicellular green algae evolved considerably faster than its homologs in multicellular angiosperms suggests that multiple environments, including multiple large subunit partners in angiosperms, constrain the evolution of the small subunit.

Finally, an additional reason for greater large subunit sequence divergence is partial redundancy. Akihiro et al. (2005) and Crevillen et al. (2005) showed that expression patterns of some large subunits partially overlap. Even though genetic studies showed that one large subunit (APL1) is of major importance in Arabidopsis leaves (Lin et al., 1988), other large subunits (APL3 and APL4) may also contribute to leaf AGPase activity. Indeed, trehalose-induced APL3 expression complemented an APL1 knockout mutant (adg2) (Fritzius et al., 2001). Consistent with this model, the angiosperm large subunits evolve slightly faster than the large subunit from the unicellular green algae. While there is one major large subunit in the investigated unicellular algae, there are multiple large subunits in the angiosperms. This may accelerate large subunit evolution due to partial redundancy.

Overall, this study shows that the higher conservation in the small subunit than the large subunit of AGPase is due to greater evolutionary constraints on the small subunit. We also show that the large and the small subunits of AGPase are equally tolerant to nonsynonymous mutations with respect to enzyme activity when expressed with a single partner in a common cellular environment. This implies that both subunits are equally important for the structure and function of AGPase. While we cannot formally exclude the possibility that subtle (and as yet unknown) differences distinguish the mechanism AGPase catalysis of the maize enzyme in E. coli versus in the maize endosperm, we suggest that the difference in evolutionary rates of sequence divergence in angiosperms is caused by the different number of genes encoding the small and large subunits of AGPase. The detailed analyses in focused studies like this one complement large-scale bioinformatics studies (e.g., Wolf et al., 2006) and suggest that comprehensive examination of other paralogs exhibiting different degrees of sequence conservation will prove useful.


Random Mutagenesis

Mutations were introduced into Sh2 and Bt2 by PCR random mutagenesis (GeneMorph II EZClone domain mutagenesis kit; Stratagene). A mixture of nonbiased, error-prone DNA polymerases was used to introduce point mutations. Wild-type Sh2 and Bt2 coding sequences in pMONcSh2 and pMONcBt2 (Giroux et al., 1996), respectively, were used as templates for PCR. Two pairs of primers (Sh2, 5′-GAAGGAGATATATCCATGG-3′ and 5′-GGATCCCCGGGTACCGAGCTC-3′; Bt2, 5′-GAAGGAGATATATCCATGG-3′ and 5′-GTTGATATCTGAATTCGAGCTC-3′) flanking Sh2 and Bt2 were used for error-prone PCR. Mutant sh2 clones produced by PCR were subcloned into vector pMONcSh2 according to Stratagene protocols. pMONcSh2 was then used to transform Escherichia coli strain AC70R1-504 that contained wild-type Bt2 in the compatible vector pMONcBt2. Mutant bt2 clones produced by PCR were subcloned into vector pMONcBt2. pMONcBt2 was then used to transform E. coli strain AC70R1-504 that contained wild-type Sh2 in the compatible vector pMONcSh2.

Bacterial Expression System

A bacterial expression system (Iglesias et al., 1993) allowed us to randomly mutagenize maize (Zea mays) endosperm AGPase genes and score AGPase activity in an efficient way. The E. coli system is ideal for analyzing plant AGPases for two reasons. First, compared with the enzyme from the host tissue, E. coli–expressed plant AGPases exhibit virtually identical kinetic and allosteric properties (Iglesias et al., 1993; Boehlein et al., 2005). Second, mutations affecting both kinetic and allosteric properties and heat stability of plant AGPases were isolated in E. coli (Greene et al., 1996; Greene and Hannah, 1998; Laughlin et al., 1998; Kavakli et al., 2001; Cross et al., 2004; Meyer et al., 2004; Boehlein et al., 2005).

Glycogen Detection

Glycogen synthesis was detected by production of brown staining colonies following a 2-min exposure to iodine vapors. E. coli cells were grown on Luria broth media in the presence of 75 μg/mL spectinomycin, 50 μg/mL kanamycin, and 2% (w/v) glucose. Iodine staining provided a fast and convenient means to identify active and inactive AGPases. Colonies with inactive AGPase produced no color following contact with iodine vapors, while active AGPase produced glycogen and, in turn, brown staining with iodine.

Glycogen Quantitation

Glycogen quantitation of AGPase mutants was performed by phenol reaction (Hanson and Phillips, 1981). In brief, glycogen was extracted from E. coli cells (OD600 = 2.0) grown in Luria broth containing 2% (w/v) glucose by boiling for 3 h in 50% (w/v) KOH. Glycogen was then precipitated by adding ethanol to 70% (v/v) and centrifuging at 10,000g for 10 min. After pellet drying, 200 μL of deionized water, 200 μL of 5% (w/v) phenol, and 1 mL of concentrated sulfuric acid were added. Glycogen was estimated by the absorbance at 488 nm.

DNA Sequencing

Five hundred and seventy six sh2 and bt2 mutants were double-pass sequenced by the Genome Sequencing Services Laboratory of the Interdisciplinary Center for Biotechnology Research at the University of Florida. Data analysis was performed by Bioedit software (Hall, 1999).

Classification of Amino Acid Changes (Conservative and Nonconservative)

BLOSUM62 matrix was used for determining conservative and nonconservative amino acid changes (Henikoff and Henikoff, 1992). Amino acid changes that were given a positive or neutral score by the BLOSUM62 matrix were considered conservative (Cargill et al., 1999). Amino acid changes that were given a negative score were considered nonconservative.

Sequence Retrieval and Alignment

We retrieved all full-length plant large and small subunits from the National Center for Biotechnology Information database. The AGPase isoforms of Populus trichocarpa were obtained from the U.S. Department of Energy Joint Genome Institute (http://genome.jgi-psf.org/Poptr1_1/Poptr1_1.home.html) by a tBLASTn search using BT2 and SH2 as queries. The AGPase isoforms of Chlamydomonas reinhardtii, Ostreococcus tauri, and Ostreococcus lucimarinus were obtained from the Joint Genome Institute (http://genome.jgi-psf.org/Chlre3/Chlre3.home.html, http://genome.jgi-psf.org/Ostta4/Ostta4.home.html, and http://genome.jgi-psf.org/Ost9901_3/Ost9901_3.home.html) either as annotated sequences or by tBLASTn searches. Alignments of 31 full-length small subunits sequences and 43 full-length large subunits sequences were obtained using the BIOEDIT software (Hall, 1999) with BLOSUM62 matrix followed by manual inspection of the resulting alignments. The poorly aligned N termini (~50 amino acids) of both the large and the small subunits were excluded from alignment. Pairwise identity was estimated using BIOEDIT and the IDENTITY matrix.

Phylogenetic Analysis

Maximum likelihood (ML) estimates of gene trees were obtained using nucleotide data and GARLI (Genetic Algorithm for Rapid Likelihood Inference) software (Zwickl, 2006). These analyses used the default General Time Reversible model of sequence evolution (Yang, 1994) and accommodating among-sites rate heterogeneity by assuming some sites are invariant and the rates at the remaining sites are Γ (gamma) distributed (Gu et al., 1995). Bootstrap values (Felsenstein, 1985) were also calculated by GARLI using 100 replicates. Branch lengths based on amino acid changes were estimated using AAML, which is part of the PAML (Phylogenetic Analysis by Maximum Likelihood) package (Yang, 1997), using the Dayhoff (PAM) empirical model (Dayhoff et al., 1978) with Γ distributed rates. Nonsynonymous to synonymous substitution ratios (ω) for the large and small subunit coding sequences were estimated using CODEML, distributed with the PAML package. ML estimates of phylogenetic trees using amino acid sequences were obtained using the rapid search algorithm implemented in PhyML (Guindon and Gascuel, 2003) and the Dayhoff model with Γ distributed rates, and the same program was used to estimate bootstrap support using 100 replicates.

BT2 Purification

Formaldehyde cross-linking of proteins associated in maize endosperm cells was done according to a modified protocol described by Rohila et al. (2004). In brief, developing maize endosperm harvested 18 d after pollination and stored at −80°C were cut into small pieces and treated with 2% (w/v) formaldehyde in ice-cold PBS buffer for 4 h. Cross-linking was quenched by addition of 2 M glycine for 1 h. Total proteins were then extracted as described by Boehlein et al. (2005). A BT2 monoclonal antibody column described by Boehlein et al. (2005) was used to purify BT2 along with associated proteins. The purification products were run on a 7.5% SDS polyacrylamide gel (+DTT) and visualized by staining with Coomassie Brilliant Blue (Laemmli, 1970).

Construction of the pMONcAgplemzm Expression Vector

To construct pMONcAgplemzm, a pUC-based cDNA clone, originally termed pcAgp1 (Giroux and Hannah, 1994), of the AGP large subunit from maize embryo was used as template. To remove an SstI site from this clone, two PCR amplifications resulting in two overlapping products were performed. The resulting PCR products were combined and used as template to PCR amplify the complete Agplemzm coding sequence without the transit peptide. Length of the transit peptide (47 amino acids) was determined using the ChloroP 1.1 server. A primer was designed to place the ATG start codon and GCC alanine codon (to form an NcoI site for cloning) 5′ of base 330 of the Agplemzm sequence. The primers designed to amplify the 5′ and 3′ ends of the insert were GGGGCCATGGCCTTCAGTGCAAGGGGTGCTGTG and CCCCGAGCTCACTATATGACGGTGCCGTCCTTG, respectively. The primers designed for mutation of the internal SstI site were CGCCTAAACTCTGGATGCGAACTCAAGAATACCATGATGATGGG and its reverse complement. Vent DNA polymerase was used in the first amplification, and Taq DNA polymerase was used in the second amplification. Products were purified using Millipore's Montage-PCR filter units after each amplification. The final purified product was digested with NcoI and SstI and ligated into the pMONcSh2 vector, and the resulting plasmid insert was confirmed by sequence analysis.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers S24991, P30523, AAQ14870, AAK69628, AAK39640, CAA88449, AAO16183, XP_481806, AAK27313, AAK27721, AAK27720, CAB89863, AAK27684, CAA65540, CAA65539, CAA54259, CAA54260, BAC66693, P23509, CAA58473, CAB01912, AAB00482, CAA58475, AAF66434, AAF66435, AAB09585, AAB91466, AAS00541, AAD56041, CAA55515, AAB91462, AAF75832, AAS88879, CAA79980, P55241, CAA86227, AAB94012, CAA47626, AAC49729, BAA23490, AAB38781, AAT78793, NP_911710, AAK27718, AAK27719, AAK27685, CAA65541, BAC66692, CAA53741, CAA52917, CAA43490, AAC21562, AAC49941, AAC49943, AAC49942, AAD56405, AAF66436, AAM14190, AAM20291, CAA77173, BAA76362, AAB91468, AAB91467, AAS00542, AAD56042, CAA55516, AAB91463, AAB91464, AAM95945, and AAS88891 and in the U.S. Department of Energy Joint Genome Institute data library under identification numbers 573143, 32753, 720391, 346791, 721986, 546089, 685406, 57133, 187891, and 42209.

Supplemental Data

The following materials are available in the online version of this article.

  • Supplemental Figure 1. Alignment of Maize Endosperm Large (SH2) and Small Subunits (BT2).
  • Supplemental Figure 2. Placement of Point Mutations Created by Error-Prone PCR in Sh2 and Bt2 Clones.
  • Supplemental Figure 3. Differential Rates of Evolution between Bt2 and Zea mays 2 after Gene Duplication.
  • Supplemental Table 1. Accession Numbers of Small and Large AGPase Subunits.
  • Supplemental Table 2. Amino Acid Changes Not Tolerated in BT2 and SH2.
  • Supplemental Table 3. Distribution of Observed and Expected Positive Clones if Multiple Intragenic Mutations Function Independently.
  • Supplemental Table 4. Distribution of Positive Clones with Different Numbers of Nonsynonymous Mutations from the Heavily Mutagenized 3rd and 4th Libraries.
  • Supplemental Table 5. Comparison of AGPase Small and Large Subunits from Unicellular Green Algae.

Supplementary Material

[Supplemental Data]


We thank William Farmerie and Regina Shaw for their DNA sequencing efforts and Susan Boehlein for her help with BT2 purification. We also thank Jon Stewart, Ken Cline, Robert Ferl, Donald McCarty, Martha Wayne, Rongling Wu, George Papageorgiou, Jon Martin, and members of the Hannah Laboratory for many useful comments and discussions. This research was supported by National Science Foundation Grants IOB-0444031, DBI-0077676, DBI-0606607, and IOB-9982626 and USDA Grant 2006-35100-17220.


The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantcell.org) is: L. Curtis Hannah (ude.lfu.safi.liam@hannah).

[W]Online version contains Web-only data.



  • Akashi, H. (2001). Gene expression and molecular evolution. Curr. Opin. Genet. Dev. 11 660–666. [PubMed]
  • Akihiro, T., Mizuno, K., and Fujimura, T. (2005). Gene expression of ADP-glucose pyrophosphorylase and starch contents in rice cultured cells are cooperatively regulated by sucrose and ABA. Plant Cell Physiol. 46 937–946. [PubMed]
  • Bae, J.M., Giroux, M.J., and Hannah, L.C. (1990). Cloning and molecular characterization of the brittle-2 gene of maize. Maydica 35 317–322.
  • Baldauf, S.L., Roger, A.J., Wenk-Siefert, I., and Doolittle, W.F. (2000). A kingdom-level phylogeny of eukaryotes based on combined protein data. Science 290 972–977. [PubMed]
  • Ballicora, M.A., Dubay, J.R., Devillers, C.H., and Preiss, J. (2005). Resurrecting the ancestral enzymatic role of a modulatory subunit. J. Biol. Chem. 280 10189–10195. [PubMed]
  • Ballicora, M.A., Iglesias, A.A., and Preiss, J. (2003). ADP-glucose pyrophosphorylase, a regulatory enzyme for bacterial glycogen synthesis. Microbiol. Mol. Biol. Rev. 67 213–225. [PMC free article] [PubMed]
  • Ballicora, M.A., Laughlin, M.J., Fu, Y., Okita, T.W., Barry, G.F., and Preiss, J. (1995). Adenosine 5′-diphosphate-glucose pyrophosphorylase from potato tuber. Significance of the N-terminus of the small subunit for catalytic properties and heat stability. Plant Physiol. 109 245–251. [PMC free article] [PubMed]
  • Ben Ali, A., De Baere, R., Van der Auwera, G., De Wachter, R., and Van de Peer, Y. (2001). Phylogenetic relationships among algae based on complete large-subunit rRNA sequences. Int. J. Syst. Evol. Microbiol. 51 737–749. [PubMed]
  • Bhave, M.R., Lawrence, S., Barton, C., and Hannah, L.C. (1990). Identification and molecular characterization of shrunken-2 cDNA clones of maize. Plant Cell 2 581–588. [PMC free article] [PubMed]
  • Bloom, J.D., Drummond, D.A., Arnold, F.H., and Wilke, C.O. (2006). Structural determinants of the rate of protein evolution in yeast. Mol. Biol. Evol. 23 1751–1761. [PubMed]
  • Boehlein, S.K., Sewell, A.K., Cross, J., Stewart, J.D., and Hannah, L.C. (2005). Purification and characterization of adenosine diphosphate glucose pyrophosphorylase from maize/potato mosaics. Plant Physiol. 138 1552–1562. [PMC free article] [PubMed]
  • Braun, E.L., Fuge, E.K., Padilla, P.A., and Werner-Washburne, M. (1996). A stationary-phase gene in Saccharomyces cerevisiae is a member of a novel, highly conserved gene family. J. Bacteriol. 178 6865–6872. [PMC free article] [PubMed]
  • Burger, B.T., Cross, J., Shaw, J.R., Caren, J., Greene, T.W., Okita, T.W., and Hannah, L.C. (2003). Relative turnover numbers of maize endosperm and potato tuber ADP-glucose pyrophosphorylases in the absence and presence of 3-PGA. Planta 217 449–456. [PubMed]
  • Cargill, M., et al. (1999). Characterization of single-nucleotide polymorphisms in coding regions of human genes. Nat. Genet. 22 231–238. [PubMed]
  • Crevillen, P., Ballicora, M.A., Merida, A., Preiss, J., and Romero, J.M. (2003). The different large subunit isoforms of Arabidopsis thaliana ADP-glucose pyrophosphorylase confer distinct kinetic and regulatory properties to the heterotetrameric enzyme. J. Biol. Chem. 278 28508–28515. [PubMed]
  • Crevillen, P., Ventriglia, T., Pinto, F., Orea, A., Merida, A., and Romero, J.M. (2005). Differential pattern of expression and sugar regulation of Arabidopsis thaliana ADP-glucose pyrophosphorylase-encoding genes. J. Biol. Chem. 280 8143–8149. [PubMed]
  • Cross, J.M., Clancy, M., Shaw, J., Boehlein, S., Greene, T., Schmidt, R., Okita, T., and Hannah, L.C. (2005). A polymorphic motif in the small subunit of ADP-glucose pyrophosphorylase modulates interactions between the small and large subunits. Plant J. 41 501–511. [PubMed]
  • Cross, J.M., Clancy, M., Shaw, J., Greene, T.W., Schmidt, R.R., Okita, T.W., and Hannah, L.C. (2004). Both subunits of ADP-glucose pyrophosphorylase are regulatory. Plant Physiol. 135 137–140. [PMC free article] [PubMed]
  • Dayhoff, M.O., Schwartz, R.M., and Orcutt, B.C. (1978). A model of evolutionary change in proteins. In Atlas of Protein Sequence and Structure, Vol. 5, Suppl. 3, M.O. Dayhoff, ed (Silver Springs, MD: National Biomedical Research Foundation), pp. 345–352.
  • Dickerson, R.E. (1971). The structure of cytochrome C and the rates of molecular evolution. J. Mol. Evol. 1 26–45. [PubMed]
  • Drummond, D.A., Bloom, J.D., Adami, C., Wilke, C.O., and Arnold, F.H. (2005). Why highly expressed proteins evolve slowly. Proc. Natl. Acad. Sci. USA 102 14338–14343. [PMC free article] [PubMed]
  • Duret, L., and Mouchiroud, D. (2000). Determinants of substitution rates in mammalian genes, expression pattern affects selection intensity but not mutation rate. Mol. Biol. Evol. 17 68–74. [PubMed]
  • Felsenstein, J. (1981). Evolutionary trees from DNA sequences: A maximum likelihood approach. J. Mol. Evol. 17 368–376. [PubMed]
  • Felsenstein, J. (1985). Confidence-limits on phylogenies: An approach using the bootstrap. Evolution Int. J. Org. Evolution 39 783–791.
  • Fraser, H.B., Hirsh, A.E., Steinmetz, L.M., Scharfe, C., and Feldman, M.W. (2002). Evolutionary rate in the protein interaction network. Science 296 750–752. [PubMed]
  • Fritzius, T., Aeschbacher, R., Wiemken, A., and Wingler, A. (2001). Induction of ApL3 expression by trehalose complements the starch-deficient Arabidopsis mutant adg2-1 lacking ApL1, the large subunit of ADP-glucose pyrophosphorylase. Plant Physiol. 126 883–889. [PMC free article] [PubMed]
  • Frueauf, J.B., Ballicora, M.A., and Preiss, J. (2003). ADP-glucose pyrophosphorylase from potato tuber, site-directed mutagenesis of homologous aspartic acid residues in the small and large subunits. Plant J. 33 503–511. [PubMed]
  • Fu, Y., Ballicora, M.A., and Preiss, J. (1998). Mutagenesis of the glucose-1-phosphate-binding site of potato tuber ADP-glucose pyrophosphorylase. Plant Physiol. 117 989–996. [PMC free article] [PubMed]
  • Fuchs, R. L. (1977). Purification and Characterization of ADP-Glucose Pyrophosphorylase A from Maize Endosperm. PhD dissertation (College Station, TX: Texas A&M University).
  • Giroux, M., Smith-White, B., Gilmore, V., Hannah, L.C., and Preiss, J. (1995). The large subunit of the embryo isoform of ADP glucose pyrophosphorylase from maize. Plant Physiol. 108 1333–1334. [PMC free article] [PubMed]
  • Giroux, M.J., and Hannah, L.C. (1994). ADP-glucose pyrophosphorylase in shrunken-2 and brittle-2 mutants of maize. Mol. Gen. Genet. 243 400–408. [PubMed]
  • Giroux, M.J., Shaw, J., Barry, G., Cobb, B.G., Greene, T.W., Okita, T.W., and Hannah, L.C. (1996). A single mutation that increases maize seed weight. Proc. Natl. Acad. Sci. USA 93 5824–5829. [PMC free article] [PubMed]
  • Goldman, N. (1993). Statistical tests of models of DNA substitution. J. Mol. Evol. 36 182–198. [PubMed]
  • Greene, T.W., Chantler, S.E., Kahn, M.L., Barry, G.F., Preiss, J., and Okita, T.W. (1996). Mutagenesis of the potato ADPglucose pyrophosphorylase and characterization of an allosteric mutant defective in 3-phosphoglycerate activation. Proc. Natl. Acad. Sci. USA 93 1509–1513. [PMC free article] [PubMed]
  • Greene, T.W., and Hannah, L.C. (1998). Enhanced stability of maize endosperm ADP-glucose pyrophosphorylase is gained through mutants that alter subunit interactions. Proc. Natl. Acad. Sci. USA 95 13342–13347. [PMC free article] [PubMed]
  • Gu, X., Fu, Y.X., and Li, W.H. (1995). Maximum likelihood estimation of the heterogeneity of substitution rate among nucleotide sites. Mol. Biol. Evol. 12 546–557. [PubMed]
  • Guindon, S., and Gascuel, O. (2003). A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst. Biol. 52 696–704. [PubMed]
  • Guo, H.H., Choe, J., and Loeb, L.A. (2004). Protein tolerance to random amino acid change. Proc. Natl. Acad. Sci. USA 101 9205–9210. [PMC free article] [PubMed]
  • Hahn, M.W., and Kern, A.D. (2005). Comparative genomics of centrality and essentiality in three eukaryotic protein-interaction networks. Mol. Biol. Evol. 22 803–806. [PubMed]
  • Hall, D.B., and Struhl, K. (2002). The VP16 activation domain interacts with multiple transcriptional components as determined by protein-protein cross-linking in vivo. J. Biol. Chem. 277 46043–46050. [PubMed]
  • Hall, T.A. (1999). BioEdit, a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41 95–98.
  • Hannah, L.C. (2005). Starch synthesis in the maize endosperm. Maydica 50 497–506.
  • Hannah, L.C., and Nelson, O.E., Jr. (1976). Characterization of ADP-glucose pyrophosphorylase from shrunken-2 and brittle-2 mutants of maize. Biochem. Genet. 14 547–560. [PubMed]
  • Hannah, L.C., Shaw, J.R., Giroux, M.J., Reyss, A., Prioul, J.L., Bae, J.M., and Lee, J.Y. (2001). Maize genes encoding the small subunit of ADP-glucose pyrophosphorylase. Plant Physiol. 127 173–183. [PMC free article] [PubMed]
  • Hanson, R.S., and Phillips, J.A. (1981). Chemical composition. In Manual of Methods for General Bacteriology, P. Gerhandt, R.G.E. Murray, R.N. Costilow, E.W. Nester, W.A. Wood, N.R. Krieg, and G.B. Phillips, eds (Washington D.C.: American Society for Microbiology), pp. 328–364.
  • Hastings, K.E. (1996). Strong evolutionary conservation of broadly expressed protein isoforms in the troponin I gene family and other vertebrate gene families. J. Mol. Evol. 42 631–640. [PubMed]
  • Henikoff, S., and Henikoff, J.G. (1992). Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci. USA 89 10915–10919. [PMC free article] [PubMed]
  • Hileman, L.C., and Baum, D.A. (2003). Why do paralogs persist? Molecular evolution of CYCLOIDEA and related floral symmetry genes in Antirrhineae (Veronicaceae). Mol. Biol. Evol. 20 591–600. [PubMed]
  • Hirsh, A.E., and Fraser, H.B. (2001). Protein dispensability and rate of evolution. Nature 411 1046–1049. [PubMed]
  • Huelsenbeck, J.P., and Rannala, B. (1997). Phylogenetic methods come of age: Testing hypotheses in an evolutionary context. Science 276 227–232. [PubMed]
  • Hurst, L.D., and Smith, N.G.C. (1999). Do essential genes evolve slowly? Curr. Biol. 9 747–750. [PubMed]
  • Hwang, S.K., Hamada, S., and Okita, T.W. (2006). ATP binding site in the plant ADP-glucose pyrophosphorylase large subunit. FEBS Lett. 580 6741–6748. [PubMed]
  • Hwang, S.K., Hamada, S., and Okita, T.W. (2007). Catalytic implications of the higher plant ADP-glucose pyrophosphorylase large subunit. Phytochemistry 68 464–477. [PubMed]
  • Iglesias, A., Barry, G.F., Meyer, C., Bloksberg, L., Nakata, P., Greene, T., Laughlin, M.J., Okita, T.W., Kishore, G.M., and Preiss, J. (1993). Expression of the potato tuber ADP-glucose pyrophosphorylase in Escherichia coli. J. Biol. Chem. 268 1081–1086. [PubMed]
  • Johnson, P.E., Patron, N.J., Bottrill, A.R., Dinges, J.R., Fahy, B.F., Parker, M.L., Waite, D.N., and Denyer, K. (2003). A low-starch barley mutant, riso 16, lacking the cytosolic small subunit of ADP-glucose pyrophosphorylase, reveals the importance of the cytosolic isoform and the identity of the plastidial small subunit. Plant Physiol. 131 684–696. [PMC free article] [PubMed]
  • Jordan, I.K., Rogozin, I.B., Wolf, Y.I., and Koonin, E.V. (2002). Essential genes are more evolutionarily conserved than are nonessential genes in bacteria. Genome Res. 12 962–968. [PMC free article] [PubMed]
  • Jordan, I.K., Wolf, Y.I., and Koonin, E.V. (2003). No simple dependence between protein evolution rate and the number of protein-protein interactions: Only the most prolific interactors tend to evolve slowly. BMC Evol. Biol. 3 1. [PMC free article] [PubMed]
  • Kavakli, I.H., Greene, T.W., Salamone, P.R., Choi, S.B., and Okita, T.W. (2001). Investigation of subunit function in ADP-glucose pyrophosphorylase. Biochem. Biophys. Res. Commun. 281 783–787. [PubMed]
  • Kimura, M. (1983). The Neutral Theory of Molecular Evolution. (Cambridge, UK: Cambridge University Press).
  • La Cognata, U., Willmitzer, L., and Müller-Röber, B. (1995). Molecular cloning and characterization of novel isoforms of potato ADP-glucose pyrophosphorylase. Mol. Gen. Genet. 246 538–548. [PubMed]
  • Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 680–685. [PubMed]
  • Laughlin, M.J., Payne, J.W., and Okita, T.W. (1998). Substrate binding mutants of the higher plant ADP-glucose pyrophosphorylase. Phytochemistry 47 621–629. [PubMed]
  • Li, W.H. (1997). Molecular Evolution. (Sunderland, MA: Sinauer Associates).
  • Liao, B.-Y., Scott, N.M., and Zhang, J. (2006). Impacts of gene essentiality, expression pattern, and gene compactness on the evolutionary rate of mammalian proteins. Mol. Biol. Evol. 23 2072–2080. [PubMed]
  • Lin, T.P., Caspar, T., Somerville, C.R., and Preiss, J. (1988). A starch deficient mutant of Arabidopsis thaliana with low ADPglucose pyrophosphorylase activity lacks one of the two subunits of the enzyme. Plant Physiol. 88 1175–1181. [PMC free article] [PubMed]
  • Marin, I., Fares, M.A., Gonzalez-Candelas, F., Barrio, E., and Moya, A. (2001). Detecting changes in the functional constraints of paralogous genes. J. Mol. Evol. 52 17–28. [PubMed]
  • Meyer, F.D., Smidansky, E.D., Beecher, B., Greene, T.W., and Giroux, M.J. (2004). The maize Sh2r6hs ADP-glucose pyrophosphorylase (AGP) large subunit confers enhanced AGP properties in transgenic wheat (Triticum aestivum). Plant Sci. 167 899–911.
  • Nei, M. (1987). Molecular Evolutionary Genetics. (New York: Columbia University Press).
  • Ohdan, T., Francisco, P.B., Jr., Sawada, T., Hirose, T., Terao, T., Satoh, H., and Nakamura, Y. (2005). Expression profiling of genes involved in starch synthesis in sink and source organs of rice. J. Exp. Bot. 56 3229–3244. [PubMed]
  • Pal, C., Papp, B., and Hurst, L.D. (2001). Highly expressed genes in yeast evolve slowly. Genetics 158 927–931. [PMC free article] [PubMed]
  • Park, S.W., and Chung, W.I. (1998). Molecular cloning and organ-specific expression of three isoforms of tomato ADP-glucose pyrophosphorylase gene. Gene 206 215–221. [PubMed]
  • Prioul, J.L., Jeannette, E., Reyss, A., Gregory, N., Giroux, M., Hannah, L.C., and Causse, M. (1994). Expression of ADP-glucose pyrophosphorylase in maize (Zea mays L.) grain and source leaf during grain filling. Plant Physiol. 104 179–187. [PMC free article] [PubMed]
  • Ral, J.P., et al. (2004). Starch division and partitioning. A mechanism for granule propagation and maintenance in the picophytoplanktonic green alga Ostreococcus tauri. Plant Physiol. 136 3333–3340. [PMC free article] [PubMed]
  • Rocha, E.P.C., and Danchin, A. (2004). An analysis of determinants of amino acids substitution rates in bacterial proteins. Mol. Biol. Evol. 21 108–116. [PubMed]
  • Rohila, J.S., Chen, M., Cerny, R., and Fromm, M.E. (2004). Improved tandem affinity purification tag and methods for isolation of protein heterocomplexes from plants. Plant J. 38 172–181. [PubMed]
  • Rosti, S., Rudi, H., Rudi, K., Opsahl-Sorteberg, H.G., Fahy, B., and Denyer, K. (2006). The gene encoding the cytosolic small subunit of ADP-glucose pyrophosphorylase in barley endosperm also encodes the major plastidial small subunit in the leaves. J. Exp. Bot. 57 3619–3626. [PubMed]
  • Schmidt, T.R., Jaradat, S.A., Goodman, M., Lomax, M.I., and Grossman, L.I. (1997). Molecular evolution of cytochrome c oxidase: Rate variation among subunit VIa isoforms. Mol. Biol. Evol. 14 595–601. [PubMed]
  • Smith-White, B.J., and Preiss, J. (1992). Comparison of proteins of ADP-glucose pyrophosphorylase from diverse sources. J. Mol. Evol. 34 449–464. [PubMed]
  • Thorbjørnsen, T., Villand, P., Denyer, K., Olsen, O., and Smith, A.M. (1996). Distinct isoforms of ADPglucose pyrophosphorylase occur inside and outside the amyloplast in barley endosperm. Plant J. 10 243–250.
  • Tuskan, G.A., et al. (2006). The genome of black cottonwood, Populus trichocarpa (Torr. & Gray). Science 313 1596–1604. [PubMed]
  • Van de Peer, Y., Rensing, S.A., Maier, U.G., and De Wachter, R. (1996). Substitution rate calibration of small subunit ribosomal RNA identifies chlorarachniophyte endosymbionts as remnants of green algae. Proc. Natl. Acad. Sci. USA 93 7732–7736. [PMC free article] [PubMed]
  • Wall, D.P., Hirsh, A.E., Fraser, H.B., Kumm, J., Giaever, G., Eisen, M.B., and Feldman, M.W. (2005). Functional genomic analysis of the rates of protein evolution. Proc. Natl. Acad. Sci. USA 102 5483–5488. [PMC free article] [PubMed]
  • Wilson, A.C., Carlson, S.S., and White, T.J. (1977). Biochemical evolution. Annu. Rev. Biochem. 46 573–639. [PubMed]
  • Wolf, Y.I., Carmel, L., and Koonin, E.V. (2006). Unifying measures of gene function and evolution. Proc. R. Soc. Lond. B Biol. Sci. 273 1507–1515. [PMC free article] [PubMed]
  • Wright, S.I., Yau, C.B., Looseley, M., and Meyers, B.C. (2004). Effects of gene expression on molecular evolution in Arabidopsis thaliana and Arabidopsis lyrata. Mol. Biol. Evol. 21 1719–1726. [PubMed]
  • Yang, Z. (1994). Estimating the pattern of nucleotide substitution. J. Mol. Evol. 39 105–111. [PubMed]
  • Yang, Z. (1997). PAML: A program package for phylogenetic analysis by maximum likelihood. Comput. Appl. Biosci. 13 555–556. [PubMed]
  • Yang, Z., Goldman, N., and Friday, A. (1995). Maximum likelihood trees from DNA sequences: a peculiar statistical estimation problem. Syst. Biol. 44 384–399.
  • Zabawinski, C., Van den Koornhuyse, N., D'Hulst, C., Schlichting, R., Giersch, C., Delrue, B., Lacroix, J.M., Preiss, J., and Ball, S. (2001). Starchless mutants of Chlamydomonas reinhardtii lack the small subunit of a heterotetrameric ADP-glucose pyrophosphorylase. J. Bacteriol. 183 1069–1077. [PMC free article] [PubMed]
  • Zuckerkandl, E. (1976). Evolutionary processes and evolutionary noise at the molecular level: I. Functional density of proteins. J. Mol. Evol. 7 167–183. [PubMed]
  • Zwickl, D.J. (2006). Genetic Algorithm Approaches for the Phylogenetic Analysis of Large Biological Sequence Datasets under the Maximum Likelihood Criterion. PhD dissertation (Austin, TX: University of Texas).

Articles from The Plant Cell are provided here courtesy of American Society of Plant Biologists
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...