Commentary A positive addition to a negative tail's tale Arno L. Greenleaf Biochemistry Department, Duke University Medical Center, Durham, NC 27710

The C-terminal repeat domain (CTD) of threonine hyperphosphorylation, this -proximal paused complexes, the largest subunit ofRNA II modification produces a shift in SDS/gel thereby allowing productive elongation. is composed of multiple repeats of the mobility. In contrast to c-Abl, a different The presence of CTD activity in consensus heptamer sequence Tyr-Ser- , c-Src, does not phos- preinitiation complexes and the apparent Pro-Thr-Ser-Pro-Ser (YSPTSPS in sin- phorylate the CTD in vitro. This addi- association of CTD kinase activity with gle-letter code) (for reviews, see refs. tional information forces a significant re- factor TFIIH are consistent 1-3); thus five of the seven residues of evaluation of our ideas about CTD phos- with this suggestion (refs. 8-10 and the each repeat are potential sites of phos- phorylation. references therein). More recent in vivo phorylation. Indeed, hyperphosphoryla- One obvious question raised by these experiments are also in general agree- tion ofthis domain has been documented findings concerns the identity of the ty- ment with this overall picture (11). It in organisms from yeast to human (one rosine kinase(s) that phosphorylates the should be pointed out, however, that diagnostic for hyperphosphorylation of CTD in vivo. The previously known and despite the identification and character- the CTD is a marked mobility change of now reported properties ofc-Abl are con- ization of a number of different CTD the largest subunit in SDS gels; unphos- sistent with its playing such a role in the , the identity of the kinase(s) act- phorylated subunit "IIa" migrates with nucleus, but additional tests will be re- ing on the CTD in vivo has not been an apparent molecular mass of -215 quired to test critically this possibility. rigorously established. kDa, whereas hyperphosphorylated sub- Unpublished data cited by Baskaran et All the previously characterized CTD unit "IIo" migrates with an apparent al. (4) already suggest the involvement of kinase activities, including those found molecular mass of -240 kDa). With CTD other as-yet-unidentified tyrosine kinase associated with preinitiation complexes lengths ranging from 26 repeats in yeast activities because they observed CTD or initiation factors, are specific for ser- to 52 repeats in mammals, one could tyrosine in a c-Abl- ine or /threonine. Ifthey represent imagine that nuclear kinases en- negative 3T3 cell line (their Discussion). the activity/activities that in the above countering this domain might think they It is a bit unsettling that all previous scenario generate actively elongating had found Paradise. Until now, however, searches for CTD kinase activities, in RNA polymerase II0, then a prediction only serine/threonine kinases would extracts of fungal, animal, or plant cells, would be that elongating RNA polymer- have been thought to have this experi- have yielded serine or serine/threonine ase IIO is phosphorylated on serine and ence. This perception changes with the kinases. Of course, there might be sev- threonine but not on tyrosine. This pre- publication of the article by Baskaran, eral explanations for this situation, in- diction might be tested by using the ap- Dahmus, and Wang (4) in this issue, cluding low abundance oftyrosine kinase proach, frequently used by Dahmus and which demonstrates that the CTD is also activities or inappropriate assay condi- colleagues, of identifying productively subject to tyrosine phosphorylation. tions. The current report will certainly elongating RNA polymerase II by UV Previous work on in vivo-labeled mam- stimulate attempts to detect P-Tyr in the crosslinking to nascent transcripts (see malian RNA polymerase II had detected CTD of RNA polymerase II subunits ref. 12, for example) and coupling this CTD-derived phosphoserine and phos- from different organisms and to identify identification with analyses designed to phothreonine; no phosphotyrosine had the responsible kinases. The hint that detect P-Tyr. The outcome of such ex- been observed (5, 6). In the current work, c-Abl may be involved is exciting and, if periments will importantly shape further extra precautions were taken to preserve borne out, will provide new ideas about thinking and experimentation. If elongat- any phosphorylated tyrosine (P-Tyr) that mechanisms of abl oncogene-mediated ing RNA polymerase II contains P-Tyr, it might be present on the CTD in HeLa transformation. will then be critical to identify when and nuclei: extract preparation procedures A basic question, of course, which is by which activity the phosphates were were modified, and high levels of P-Tyr still not fully answered by experimental added. Do preinitiation complexes also phosphatase inhibitors were included. tests, is "What is the function of CTD contain CTD tyrosine kinase activity, for These steps led to detecting P-Tyr in phosphorylation?" In vitro experiments example? If so, what is the identity ofthe digestions of RNA polymerase II largest over the last few years have suggested kinase? subunit purified by immunoprecipitation the following scenario: RNA polymerase The question of P-Tyr presence on and SDS/gel electrophoresis. Under the II with unphosphorylated CTD (RNA elongating RNA polymerase II also, of conditions used approximately equal polymerase IIA) enters into preinitiation course, has implications for thinking amounts of P-Tyr and phosphorylated complexes, the CTD interacting with about possible functions of the CTD in threonine and about three to five times TFIID and possibly postinitiation phases of transcription and more phosphorylated serine were de- other factors; concurrent with initiation the regulation ofthose functions. Several tected. and/or beginning productive elongation roles have been suggested for the hyper- Further experiments reported by the CTD becomes hyperphosphorylated, phosphorylated CTD during transcript Baskaran et al. (4) demonstrate that in such that the elongating transcriptase is elongation. One suggestion is that the vitro, the CTD can be phosphorylated by RNA polymerase IIO (ref. 7 and the hyperphosphorylated CTD represents an a tyrosine kinase known to be found in references therein). This scenario leads example of an acidic polymer that facil- the nucleus, c-Abl. The c-Abl kinase can to the frequent suggestion that CTD itates passage of polymerase through nu- add up to -30 phosphates to tyrosines in phosphorylation mediates "release" of cleosomes by catalyzing the displace- the CTD and, as is the case for serine/ RNA polymerase II from preinitiation or ment of H2A/H2B dimers (13). 10896

8928/Dec:23/1939 Downloaded by guest on October 1, 2021 Commentary: Greenleaf Proc. Natl. Acad. Sci. USA 90 (1993) 10897 Another suggestion is that the phosphor- ("IIo"), which we now realize can be our attempts to understand this unusual ylated CTD is a docking or attachment caused by different kinds of hyperphos- domain and its posttranslational modifi- site for certain RNA-processing compo- phorylation (see also ref. 1). cations we are only at the beginning of a nents (14). Whatever the actual functions Another earlier experiment showed complex epic rather than near the end of of the phospho-CTD during elongation, that in a mu- a simple tale. the inventory of activities able to modu- tant strain lacking a functional CTKI late those functions may be much larger , the gene encoding the catalytic 1. Corden, J. L. (1990) Trends Biochem. than previously thought if it includes ty- subunit of a well-characterized yeast Sci. 15, 383-387. rosine kinases; the implications of this CTD kinase, very little "IIo" subunit 2. Young, R. A. (1991) Annu. Rev. Bio- eventuality would be far-reaching. could be detected by antibodies directed chem. 60, 689-715. Alternatively, if actively elongating against the serine/threonine-phosphory- 3. Dahmus, M. E. & Dynan, W. S. (1992) RNA polymerase II does not contain lated CTD; in the same strain both Ia and in Transcriptional Regulation, eds. Ya- mamoto, K. & McKnight, S. (Cold P-Tyr, very different models would en- Iho could be detected by antibodies to a Spring Harbor Lab. Press, Plainview, sue. Could, for example, tyrosine and non-CTD portion of the largest subunit NY), Vol. 1, pp. 109-129. serine/threonine phosphorylation be mu- (16). These apparently contradictory re- 4. Baskaran, R., Dahmus, M. E. & Wang, tually exclusive events with opposite sults might conceivably be explained by J. Y. J. (1993) Proc. Natl. Acad. Sci. functional consequences? Might CTD invoking tyrosine phosphorylation. USA 90, 11167-11171. hyperphosphorylation on tyrosines ren- Namely, if the subunit migrating at the 5. Cadena, D. L. & Dahmus, M. E. (1987) der RNA polymerase II incapable of ini- IIo position and detected by the non- J. Biol. Chem. 262, 12468-12474. tiation and represent a mechanism for CTD antibodies were present because of 6. Zhang, J. & Corden, J. L. (1991) J. Biol. shutting down transcription in part or hyperphosphorylation on tyrosine, it Chem. 266, 2290-2296. entirely at certain critical points in devel- would not have been detected by the 7. Chesnut, J. D., Stephens, J. H. & Dah- antibodies used be- mus, M. E. (1992) J. Biol. Chem. 267, opment or in the ? anti-phospho-CTD 10500-10506. In addition to the above questions and cause they were raised against the ser- 8. Gileadi, O., Feaver, W. J. & Komberg, speculations, this current report also in- ine/threonine-phosphorylated CTD. R. D. (1992) Science 257, 1389-1392. vites a reevaluation of previous experi- This explanation should be subject to 9. Lu, H., Zawel, L., Fisher, L., Egly, ments. For instance, because hyperphos- relatively easy experimental assessment. J. M. & Reinberg, D. (1992) Nature phorylation on tyrosines can apparently Finally, a recent report revealed that (London) 358, 641-645. cause the RNA polymerase Ila -I Io the CTD in mammalian RNA polymerase 10. Serizawa, H., Conaway, R. C. & Con- mobility shift, previous experiments that II can be modified, not only by addition away, J. W. (1992) Proc. Natl. Acad. monitored the ratio of subunit forms mi- of phosphate groups but also by the ad- Sci. USA 89, 7476-7480. a means to of 0-linked sugars, specifically 11. Weeks, J. R., Hardin, S. E., Shen, J., grating as "IIa" and "ITo" as dition Lee, J. M. & Greenleaf, A. L. (1993) measure CTD phosphorylation state in O-GlcNAc (17). In that work CTD gly- Dev., in press. crude extracts or in vivo may have been cosylation and phosphorylation were 12. Bartholomew, B., Dahmus, M. E. & influenced in unappreciated ways by ty- found to be mutually exclusive events. Meares, C. F. (1986) J. Biol. Chem. 261, rosine phosphorylation. For example, at- However, because P-Tyr may not have 14226-14231. tempts to observe differences in the Ila/ been preserved in the preparations of 13. Hansen, J. C. & Ausio, J. (1992) Trends IIo ratio as a function of position in the RNA polymerase II analyzed, the rela- Biochem. Sci. 17, 187-191. cell cycle failed to reveal any cell-cycle- tionship between glycosylation and tyro- 14. Greenleaf, A. L. (1993) Trends Biochem. dependent changes (ref. 15; S. Hardin sine phosphorylation is actually not yet Sci. 18, 117-119. clear. 15. Kolodziej, P. A., Woychik, N., Liao, and A.L.G., unpublished work). How- S. M. & Young, R. A. (1990) Mol. Cell. ever, had the relative levels of serine/ Overall, the unexpected findings pre- Biol. 10, 1915-1920. threonine vs. tyrosine phosphorylation sented in the paper by Baskaran et al. (4) 16. Lee, J. M. & Greenleaf, A. L. (1991) actually changed, the qualitatively differ- emphasize how little we really know Gene Exp. 1, 149-167. ent phosphorylation state could have about the functions ofthe CTD and about 17. Kelly, W. G., Dahmus, M. E. & Hart, been obscured by the relatively constant mechanisms and consequences of its G. W. (1993) J. Biol. Chem. 268, 10416- total amount of slower migrating subunit phosphorylation. They suggest that in 10424. Downloaded by guest on October 1, 2021