Functional unit of the RNA polymerase II C-terminal domain lies within heptapeptide pairs

Abstract

Unlike all other RNA polymerases, the largest subunit (RPB1) of eukaryotic DNA-dependent RNA polymerase II (RNAP II) has a C-terminal domain (CTD) comprising tandemly repeated heptapeptides with the consensus sequence Y-S-P-T-S-P-S. The tandem structure, heptad consensus, and most key functions of the CTD are conserved between yeast and mammals. In fact, all metazoans, fungi, and green plants examined to date, as well as the nearest protistan relatives of these multicellular groups, contain a tandemly repeated CTD. In contrast, the RNAP II largest subunits from many other eukaryotic organisms have a highly degenerate C terminus or show no semblance of the CTD whatsoever. The reasons for intense stabilizing selection on CTD structure in certain eukaryotes, and its apparent absence in others, are unknown. Here we demonstrate, through in vivo genetic complementation, that the essential functional unit of the yeast CTD is contained within pairs of heptapeptides. Insertion of a single alanine residue between diheptads has little phenotypic effect, while increasing the distance between diheptads produces a mostly quantitative effect on yeast cell growth. We further explore structural constraints on the CTD within an evolutionary context and propose selective mechanisms that could maintain a global tandem structure across hundreds of millions of years of eukaryotic evolution.

FIG. 1.

Sequences used in genetic complementation for the CTD in yeast. (A) Red algal sequences used to replace the yeast CTD. pYBH1 and pYBH2 were recovered from B. hamifera using PCR linkers engineered with AvaI sites for directional cloning. pYBH1 includes the entire sequence present inserted in place of the 26 yeast WT CTD repeats, whereas pYBH2 includes only the sequence enclosed in the left-hand bracket. pYGV1 and -2 were isolated in the same manner from G. vacuolata. None of these sequences complemented CTD function. pYGV2.2 was made by concatenating two copies of the pYGV2 insert. Right-hand brackets on pYGV sequences represent the presumably functional diheptad units present in these constructs. (B and C) Results of the plasmid shuffle assay in yeast using red algal CTD mutants. A positive control transformed with the pY1 shuttle vector, without substitution for the WT CTD sequence, is at the top of each panel, and a negative control representing yeast cells transformed with a LEU2 plasmid containing no copy of RPB1 is at the bottom. The left-hand panel shows original transformants containing both WT and mutant CTD; subsequent panels show replica plating on 5-FOA, which forces yeast cells to lose the URA3-linked copy of RPB1 with WT CTD. The two right-hand panels show cells first acclimated to 5-FOA selection at 30°C and then replica plated on YEPD (without 5-FOA) at high and low temperatures. No residual negative control is visible, due to its loss during the first round of 5-FOA selection at 30°C. Although cells from the replica transfer were visible at 15°C, subsequent plating on YEPD at 30°C demonstrated that they were inviable.

FIG. 2.

Full sequences of artificial CTD sequences used in yeast transformations, along with subcloning strategy for constructing RPB1 genes with mutated CTDs. Octadapeptide constructs with shorter artificial CTDs were screened, but all proved inviable. Therefore, only the longest sequences are included here and in our documented transformation results (see Fig. 4).

FIG. 3.

Phylogenetic tree recovered though Bayesian inference (15) on an alignment of inferred amino acid sequences of RPB1 regions A to H (17) from 31 exemplars, representing 16 lineages of eukaryotes. Maximum-likelihood branch lengths were calculated with TreePuzzle 50 (25). Both analyses applied a discrete estimate of 1 invariable + 4 γ-distributed rates among sites and a Jones, Taylor, and Thornton model for probabilities of changes among amino acids. RPB1 genes from organisms in bold encode a clear set of tandemly repeated C-terminal heptapeptides. The arrow designates an internode on the tree, with a 98% Bayesian confidence level (calculated from 20,000 sampled trees), that divides eukaryotes into two distinct groups. To the right, enclosed in a dashed box, is a CTD clade in which all organisms sampled to date contain a canonical CTD. In contrast, none of the taxa to the left of this divide has a clearly canonical CTD. Although several of these sequences have noncanonical heptad repeats, most have no indication of a tandem heptad structure whatsoever. The RPB1 sequence alignment is available at the website http://personal.ecu.edu/stillerj/rpb1aln.htm. Also, see reference for a thorough phylogenetic treatment of RPB1 sequences and the CTD clade.

FIG. 4.

Growth data for C-terminal constructs comprising diheptad repeats interrupted by Ala residues (see Fig. 2 for specific sequences of each construct). (A) Artificial constructs made by concatenating typical yeast heptads interrupted by Ala residues. Subscripts indicate diheptads as well as the number of repeat units in the construct. Transformants were assayed for the ability to grow directly under 5-FOA selection and on less-stringent media after acclimation to 5-FOA. A pair of bold pluses (++) indicates a growth rate close to that of the WT CTD under the respective plating conditions, a single bold + indicates an intermediate growth phenotype, and a + not in bold indicates an extremely slow growth phenotype. (B) Titration of liquid cultures grown for 48 h in SC plus 5-FOA medium. Cultures all were inoculated with 10⁴ cells at time zero and grown in 10 ml of medium at room temperature on an orbital rotator. Cultures were incubated for 2 days to assure that some growth would be observed in pYD5A transformants; consequently, both WT CTD and pYDA9 cultures approached stationary phase when plated and showed a slightly smaller difference in apparent growth rate than was obtained by quantitative measurements. (C) Quantitative comparison of growth rates of WT and various diheptad transformants in YEPD medium. The pYD5A cold temperature slow-growth phenotype is extreme; even on YEPD medium, plated cells did not become visible for several weeks. In general, slow-growth phenotypes were more exaggerated on selective medium (plate B) than on complete (plate C), indicative of a decline in the ability to adapt to some complex transcription requirements, as seen previously in progressively shorter CTD truncation mutants (27).