Nucleic Acids

Nucleic acids are polymers of acidic monomeric subunits known as nucleotides. The nucleotides form a duplex, or double-stranded, molecule referred to as deoxyribonucleic acid (DNA) that stores genetic information within the cell. The genetic information in DNA is transferred to ribonucleic acid (RNA), monomeric forms of nucleic acids that are primarily single-stranded molecules. The three major RNA species differ in their com-position and function. These are designated ribosomal RNA (rRNA), transfer RNA (tRNA), and messenger RNA (mRNA). The rRNA represents approximately 80% of cellular RNA, tRNA approximately 15%, and mRNA 3–5%. The RNA molecules participate in the conversion of the genetic information stored in DNA into proteins that are crucial for cellular function. Nucleic Acid Building Blocks: The nucleotide components of the nucleic acids include a heterocyclic nitrogenous base, a pentose sugar, and a phosphate ( Fig. 1 ). At physiological pH, the phosphate of the nucleotide is completely ionized to the anionic form and the nitrogenous base is linked through an N-beta glycosidic bond to the 1 0 carbon of the pentose sugar. By convention, the carbon atoms of the pentose sugars are numbered 1 0 to 5 0 , with the phosphate esterified to the 5 0 carbon of the sugar. The pentose sugars, deoxyribose, and ribose, are found in DNA and RNA, respectively. These sugars differ in the absence (deoxyribose) or presence (ribose) of a hydroxyl group at the 2 0 carbon of the pentose ( Fig. 2 ). The presence of the 2 0 hydroxyl group is responsible for the instability of RNA molecules.


INTRODUCTION
Eukaryotic tRNA genes and the adenovirus genes encoding the two viral-associated RNAs (VA1 and VA2 RNAs) contain internal transcription control regions consisting of two noncontiguous DNA segments termed the A and B boxes (1)(2)(3)(4). These genes are transcribed by RNA polymerase Ill (pol IE) in conjunction with pol HI specific transcription factors (5). One of these factors, TFEIC, has been shown to bind to the intragenic promoter regions and form a stable complex which is activated for transcription by the addition of pol In and TFIB (6)(7)(8)(9)(10)(11).
Unlike yeast TFUIC (12), TFIIIC from higher eukaryotes can be separated into two components required for in vitro transcription (13)(14)(15). We have referred to these as TFIJC1 and TFHIC2 in preparations from human cells. TFMC2, binds to the B-box promoter element of tRNA and the adenovirus type 2 VAl RNA genes (13,15) and TFIICI extends the footprint over the A-box region (13). The Ad2 VAI B-box is contained in an eighteen basepair perfect inverted repeat (1). TFRIC2 binds to this region with very high affinity (Ka=2 x I10 M-1) and produces a DNase I footprint centered over the inverted repeat (16). TFIIIC2 behaves as a high molecular weight protein with a sedimentation coefficient of -18S corresponding to a globular protein of 400-500 kDa (13). Highly purified TFMC2 contains five polypeptides ranging in size from 60 to 230 kDa (17). The largest of these can be specifically UV cross-linked to the VAl B-box sequence (16,17). Yeast TFHIC (called r) is also a large protein composed of several polypeptide chains (12). However, unlike TFIIIC2, r protects both the A-box and B-box promoter elements of tRNA genes from DNase I digestion (9,18), and two polypeptides can be UV cross-linked to DNA (12). The larger of these binds specifically to a tRNA gene sequence when bound to nitrocellulose (19). Thus yeast r appears to contain a combination of the human TFIHCI plus TFIIIC2 activities. 7761 Nucleic Acids Research Volume 17 Number 19 1989 Nucleic Acids Research Limited proteolytic digestion has been a useful procedure for defining functional domains in proteins. This technique has proven to be particularly rewarding when a single or limited number of protease-sensitive regions of a protein separate protease-resistant domains. This is the case for DNA-binding proteins such as the phage lambda repressor (20), the yeast cx2 repressor (21), the adenovirus single-stranded DNA binding protein (22), the Xenopus transcription factor TFIIIA (23), and more recently, the yeast transcription factor T (24). Proteolysis of T suggests the existence of two functional domains in the high molecular weight protein: a r-A domain which interacts with A-box sequence and is sensitive to proteolysis, and a T-B domain which binds the B-box sequence and is resistant to protease digestion (24). The TFIIIC2 protein which we studied earlier had properties similar to the T-B domain of the yeast protein in that it interacts with the B-box but not the A-box promoter element. Consequently we analyzed the properties of partially proteolyzed TFUIC2 in a study analogous to the one performed with yeast r. We found that a chymotrypsinresistant fragment of TFIIIC2 yields a DNase I footprint virtually identical to that of the intact TFUIC2 providing evidence for a DNA binding domain. Regions of TFIIIC2 removed by partial proteolysis were found to be essential for forming stable DNA-protein complexes with VA1 DNA and for complementation with TFHIC 1, TFIIIB and pol III for in vitro transcription of VAI RNA.

Purification of TFIIIC2
TFHIC2 was prepared from 293 cells grown in suspension culture in spinner MEM with 5% newborn calf serum and antibiotics. TFIIIC2 used in figures 1 and 2 was purified from S-100 extracts by chromatography on PC-Il phosphocellulose and FPLC Mono Q as described (13). TFIIC2 used in figures 3 and 4 was purified from nuclear extracts by ammonium sulfate precipitation, chromatography on S-300 and affinity chromatography on a column of multimerized B-box DNA sequence as described (17).
In vitro transcription In vitro transcription was performed as described (25) using TFIIIC1 purified through Mono Q chromatography and fraction B containing TFRIB and pol Im purified by gradient elution from phosphocellulose as described (13). The 40 il reactions were incubated for 60 min at 30°C and contained 0.6 jig TFHIC1 fraction protein, 1.3 jig fraction B protein, 0.35 yg pVA1, 0. 15 Ag pUC 18 and the indicated amounts of TFIIC2 or chymotrypsin-digested TFIIIC2. In vitro transcribed RNA was extracted and analyzed by gel electrophoresis and autoradiography as described (25).
poly(dI:dC) for 30 min at 25°C. Samples were loaded on a 3.5% polyacrylamide gel (acrylamide:bisacrylamide ratio of 50:1.5) cast in 10 mM Tris-HCl, pH 7.6, 1 mM EDTA. Prerun gels were loaded while current was on and the buffer was stirred and recirculated during electrophoresis (16)

Nucleic Acids Research
The mobility of the DNA-protein complex was dramatically altered by proteolytic digestion ( fig. 1). The specific retarded band of TFUIC2-VAI DNA complex (band IH) disappeared progressively with increasing protease concentration. In the case of chymotrypsin digestion, the slow migrating complex of band HI was replaced by a discretely-migrating complex of higher mobility designated band II at 10 and 25 ng chymotrypsin (lanes 8 and 9). The amount of probe retarded in the chymotrypsin-digested complex II was diminished compared to the amount bound by undigested TFIIIC2. An intermediate complex with mobility between that of band II and band Im was observed with 5 ng chymotrypsin (lane 7). Digestion with staph.V8 protease produced a DNA-protein complex of similar mobility to chymotrypsin-digested complex H (lanes 10-14). Trypsin digestion resulted in the progressive loss of the TFIIC2-VAI complex band Ill without the occurrence of any discrete band of higher mobility (data not shown). The sharpness of chymotrypsin-digested band II suggested the existence of a fairly homogeneous, partially digested TFIHC2-VAl complex. For this reason, and because chymotrypsin hydrolysis is readily controlled with the specific inhibitor chymostatin, chymotrypsin was the protease used in further studies. The same retarded band pattern was obtained when TFRIC2 was incubated with chymotrypsin and further proteolysis inhibited with chymostatin prior to incubation with VAI DNA (see fig. 3). Chymotrypsin-cleaved TFIIIC2 still binds to the B-box region of the VA] gene For a DNA-protein complex migrating in a neutral gel, the majority of the electric charge is carried by the DNA and the protein charge does not greatly influence its migration. Consequently, the much greater mobility of the TFLIC2-VAI DNA complex most likely corresponds to a substantial loss of protein mass from the preformed complex. By analogy with other DNA binding proteins, it appeared likely that chymotrypsin digestion had released a DNA-binding domain. To test whether the chymotrypsin-digested TFIIIC2 bound to VA1 DNA similarly to undigested TFHIC2, we analyzed the DNase I footprint produced by the two protein fractions.
TFHIC2 protects a 40 bp region of the VAI DNA centered over the B-box, from nucleotide +42 to +82 from the start of transcription (13). The same protected region occurred in DNA present in the retarded complex obtained in a gel shift assay and extracted from the gel (16). In a first set of experiments, aliquots of VA1 DNA-TFHIC2 complexes were incubated with increasing amounts of chymotrypsin. Chymotrypsin digestion was arrested by addition of chymostatin and DNase I was added under the conditions of DNase I footprinting, and the products resolved on a sequencing gel ( fig. 2A). No significant change in the pattern of B-box protection was observed following chymotrypsin digestion under conditions which converted all of the TFIIC2 to the faster migrating complex H (lanes 11 and 12).
In a second type of experiment, TFIIIC2 was incubated with the same end-labeled VA1 DNA probe and digested with chymotrypsin. Following chymotrypsin digestion the samples were digested with DNase I and analyzed by the gel retardation assay. Free probe and chymotrypsin-digested TFHIC2-VAI DNA complex isolated from a retardation gel. 5 ,ug aliquots of TFIIIC2 digested with 25 ng of chymotrypsin were incubated with VAl DNA probe as in (A) and the incubation mixtures were treated with 6 ng (lanes 2,3) or 3 ng (lanes 4,5) of DNase I. The resulting complexes were analyzed by gel shift assay electrophoresis, as depicted in fig. 1. Free migrating probe (lanes 2,5) and the complex of digested TFIC2-VAI DNA present in retarded band II (lanes 3,4) were extracted and analyzed on an 8%-8M urea polyacrylamide gel. Lane 1: pBR322 Msp I markers. A schematic representation of the VAI gene with the protected region is diagramed on the right of both panels.  chymotrypsin-digested TFIIIC2 band were visualized by autoradiography of the gel, the complexes were eluted from gel slices, and the isolated DNA was denatured and analyzed on a sequencing gel ( fig.2B). The DNase I footprint associated with the chymotrypsindigested TFIIIC2 complex (lanes 3 and 4) was centered over the B-box and was very similar to the DNase I footprint observed with undigested TFIIIC2 in solution ( fig. 2A, lanes 6 8). Thus, in spite of the drastic change in the mobility of the chymotrypsin-digested TFIHC2-VAI DNA complex, there was little change in the DNase I footprint produced with the chymotrypsin-digested complex. Chyrnotrypsin-digested TFIIIC2 is inactive for in vitro transcription To determine whether chymotrypsin-digested TFIIIC2 was still transcriptionally active, TFLIIC2 was assayed in an in vitro reconstituted transcription reaction with TFIIICl1, TFIII and pol HII ( fig. 3). Equivalent gel retarding activity of undigested and chymotrypsin-digested TFHIC2 ( fig. 3A, 0.5 dt and 4 1dl, respectively) were used in the transcription reactions shown in fig. 3B. In this preparation of chymotrypsin-digested TFIIIC2, a DNA-protein complex with more rapid migration than complex H was also observed. However, this complex could not be specifically competed with B-box DNA (data not shown), and therefore represented a non-specific binding activity generated by chymotrypsin digestion. In contrast, the chymotrypsin-digested TFIIC2 complex was much less stable to competition with B-box DNA, with a dissociation half-time of less than 0.5 min.

DISCUSSION
Partially proteolyzed transcription factor IIIC2 formed a specific DNA-protein complex with the VA1 B-box promoter element which had much greater mobility in a low percentage polyacrylamide gel than the undigested TFIIIC2-DNA complex ( fig. 1). Nonetheless, proteolyzed TFHIC2 generated a DNase I footprint very similar to the footprint generated with undigested TFIIIC2 (fig.2). The much greater mobility of the digested DNA-protein complex suggests that a substantial proportion of the protein mass was removed from TFIIIC2 by partial proteolysis. Therefore, partial proteolysis resulted in the cleavage of a DNA-binding domain from the remainder of the transcription factor. We attempted to determine the size of the protease resistant DNA binding polypeptide by UV cross-linking to 32P-labeled B-box DNA. However, we were unable to detect a specifically cross-linked species even though the 230 kDa polypeptide of undigested TFIIIC2 was readily crosslinked in parallel reactions, as observed previously (16,17). The inability to cross-link chymotrypsin digested TFIIC2 was probably a result of its greatly decreased affinity. Similar studies have been performed with the well characterized lambda-repressor (21) and yeast a2 repressor (22) proteins. These proteins bind as dimers of a single polypeptide chain. Each subunit binds to symmetrical half-sites in the operator sequence. In both of these cases, the affinity of the isolated DNA binding domain for operator sites is greatly decreased compared to the affinity of the intact repressors (21,22). The decreased affinity of the isolated DNA binding domains results from a loss in cooperativity of binding to the two half-sites caused by separation of the binding domains (21,22).
The affinity of partially proteolyzed TFIHC2 for B-box sequence was also greatly diminished compared to the undigested protein. This was demonstrated by the much faster off-time of partially proteolyzed TFIIIC2 from a VAI probe when the complex was challenged with a hundred fold excess of unlabeled B-box binding sites ( fig.4). Another similarity to the situation with the lambda and a2 repressors is that the TFIIIC2 binding site, the B-box, is an inverted repeat. In the case of the VA1 gene B-box it is a perfect inverted repeat of eighteen basepairs (1). The consensus B-box from tRNA genes (5) also has an inverted repeat character. By analogy to the well studied lambda and Ce2 repressors and operators, we speculate that TFIIC2 binds to the two halves of the B-box region through symmetrical DNA binding domains which are separated by partial proteolysis.

Nucleic Acids Research
Chymotrypsin-digested TFIIIC2 is inactivated for in vitro transcriptional activity. It seems likely that portions of the TFIIC2 protein required for interacting with RNA polymerase HI and the other transcription factors required for transcription, TFIIIC 1 and TFIHB, are removed or inactivated by protease digestion. However, it also possible that the much lower stability of the chymotrypsin-igested TFIIIC2 DNA-protein complex compared to the undigested DNA-protein complex also contributes to the loss of transcriptional activity.
Marzouki et al (24) performed similar partial proteolysis studies of the yeast factor i. i protein appears to be a more complex protein than TFIIIC2 in that it protects both the A-box and B-box of tRNA genes from DNase I digestion. Partially proteolyzed i produced a specific DNA-protein complex with a tRNA gene which had much greater mobility in a low percentage polyacrylamide gel than the undigested i-DNA complex, quite similar to the situation with TFIIIC2. The partially proteolyzed i protein retained the ability to protect the B-box promoter element of a tRNA gene from DNase I digestion but lost the ability to protect the A-box region. The stability of the partially proteolyzed i-B-box DNA complex was not analyzed. Partially proteolyzed r also lost in vitro transcriptional activity.