Transcription termination by phage HK022 Nun is facilitated by COOH-terminal lysine residues.

The 109-amino acid Nun protein of prophage HK022 excludes superinfecting bacteriophage lambda by blocking transcription elongation on the lambda chromosome. Multiple interactions between Nun and the transcription elongation complex are involved in this reaction. The Nun NH(2)-terminal arginine-rich motif binds BOXB sequence in nascent lambda transcripts, whereas the COOH terminus binds RNA polymerase and contacts DNA template. Nun Trp(108) is required for interaction with DNA and transcription arrest. We analyzed the role of the adjacent Lys(106) and Lys(107) residues in the Nun reaction. Substitution of the lysine residues with arginine (K106R/K107R) had no effect on transcription arrest in vitro or in vivo. Nun K106A/K107A was partially active, whereas Nun K106D/K107D was defective in vitro and failed to exclude lambda. All mutants bound RNA polymerase and BOXB. In contrast to Nun K106R/K107R and K106A/K107A, Nun K106D/K107D did not cross-link DNA template. These results suggest that transcription arrest is facilitated by electrostatic interactions between positively charged Nun residues Lys(106) and Lys(107) and negatively charged DNA phosphate groups. These may assist intercalation of Trp(108) into template.

The 109-amino acid Nun protein of prophage HK022 excludes superinfecting bacteriophage by blocking transcription elongation on the chromosome. Multiple interactions between Nun and the transcription elongation complex are involved in this reaction. The Nun NH 2 -terminal arginine-rich motif binds BOXB sequence in nascent transcripts, whereas the COOH terminus binds RNA polymerase and contacts DNA template. Nun Trp 108 is required for interaction with DNA and transcription arrest. We analyzed the role of the adjacent Lys 106 and Lys 107 residues in the Nun reaction. Substitution of the lysine residues with arginine (K106R/ K107R) had no effect on transcription arrest in vitro or in vivo. Nun K106A/K107A was partially active, whereas Nun K106D/K107D was defective in vitro and failed to exclude . All mutants bound RNA polymerase and BOXB. In contrast to Nun K106R/K107R and K106A/ K107A, Nun K106D/K107D did not cross-link DNA template. These results suggest that transcription arrest is facilitated by electrostatic interactions between positively charged Nun residues Lys 106 and Lys 107 and negatively charged DNA phosphate groups. These may assist intercalation of Trp 108 into template.
Coliphage HK022 Nun protein excludes superinfecting bacteriophage by blocking elongation of early transcripts ( Fig.  2A; Refs. 1 and 2). Nun belongs to the arginine-rich motif (ARM) 1 family of RNA binding proteins ( Fig. 1), which includes human immunodeficiency virus proteins Tat and Rev, as well as N protein (3)(4)(5). N promotes transcription antitermination on DNA templates and is required for the growth of phage (6). During transcription elongation, the Nun NH 2 -terminal ARM motif recognizes the NUT BOXB stem-loop sequence on nascent pL and pR operon transcripts, recruiting Nun to the transcription elongation complex (TEC), and blocking attachment of N ( Fig. 2B; Refs. 7 and 8). The Nun COOH terminus binds to RNA polymerase (RNAP) (9) and contacts DNA template 7-8 bp promoter-distal to the 3Ј-OH end of the transcript (10). In vitro, Nun alone blocks elongation of the TEC (11), whereas in vivo, Nun acts in association with four Escherichia coli host factors, NusA, NusB, NusE, and NusG (11)(12)(13). Host transcription repair-coupling factor releases Nun-arrested TEC (14). The penultimate Nun tryptophan residue (Trp 108 ) mediates the interaction of Nun with the template in the TEC, possibly by intercalating into double-stranded DNA template (15).
In this article, we investigate the role of Nun residues Lys 106 and Lys 107 in transcription arrest and exclusion. We showed previously that substitution of these basic residues with aspartate (Nun K106D/K107D) strongly inhibited transcription termination in vivo (16). Nun K106D/K107D could, however, prevent translation of the N gene in an rnc host (16). This suggested that Nun K106D/K107D bound BOXB, and, along with E. coli Nus factors, blocked ribosomal access to the adjacent N ribosome-binding site (16). In this paper, we show that Nun Lys 106 and Lys 107 are not required for binding to RNAP or BOXB but are instead involved in binding to DNA template. We propose that favorable electrostatic interactions between Nun Lys 106 and Lys 107 and template phosphate groups assist intercalation of Nun Trp 108 into the DNA template.
Mutagenesis-K106D/K107D and T-Nun (Nun V96) protein mutants were constructed from the pT7NunII plasmid as described previously (16). The K106A/K107A and K106R/K107R mutations were made using QuikChange site-directed mutagenesis technique (Stratagene) from the pT7NunII plasmid. The primers used to make these mutations were as follows:  TTT TAG CAT  GAC CAC GCT GCG TTT GGG TTT CGC TGG TGA GC; K106R/  K107R110C, GCT CAC CAG CGA AAC CCA AAC AGA AGG TGG TCA  TGC TAA AAG CTT GC and GCA AGC TTT TAG CAT GAC CAC CTT  CTG TTT GGG TTT CGC TGG TGA GC; K106D/K107D110C, GCT  CAC CAG CGA AAC CCA AAC GAC GAC TGG TCA TGC TAA AAG  CTT GC and GCA AGC TTT TAG CAT GAC CAG TCG TCG TTT GGG  TTT CGC TGG TGA GC. ␤-Galactosidase and Galactokinase Assays-Cultures were grown in LB ϩ 50 g/ml of ampicillin at 32°C and shifted to 42°C for 1 h with shaking to OD 600 ϭ 0.6 for ␤-galactosidase assays (19) and for 7 h to OD 650 ϭ 0.5-0.7 for galactokinase measurements (20). The shift to 42°C inactivates the cI857 repressor and initiates transcription from the pL (␤-galactosidase assays) or pR promoter (galactokinase assays).
Efficiency of Plating-Lawns of strains were poured in top agar on LB or LB ϩ 50 g/ml ampicillin. Efficiency of plating was determined by spotting dilutions of phage and incubating overnight at 37°C.
In Vitro Transcription Assay-This assay was based on two-step, multiround transcription reaction. In the first step, open complex formation was allowed by preincubating E. coli RNAP (20 nM), DNA template (10 nM) in 50 l of transcription buffer containing 20 mM Tris acetate (pH 7.9), 60 mM potassium acetate, 2 mM magnesium acetate, 0.2 mg/ml bovine serum albumin, 1 mM DTT, and 5% glycerol at 32°C for 10 min. In the second step, transcription was initiated by addition of 5 M ATP, CTP, GTP, UTP and 1-2 Ci of [␣-32 P]ATP, and where indicated, wild-type or Nun mutant was added. The final concentration of wild type, K106A/K107A, K106R/K107R, K106D/K107D, or T-Nun in the elongation mixture was 200 nM. Transcription elongation was continued at 32°C for 20 min. The transcription reaction was then terminated by addition of a 50-l stop solution containing 375 mM sodium acetate (pH 5.2), 62.5 mM EDTA, and 1 g of yeast tRNA. Transcription products were then phenol/chloroform-extracted, ethanol-precipitated with 2.5 volumes of 100% ethanol (Sigma), and resuspended in 20 l of gel loading buffer II (Amicon). Extracted transcripts were electrophoresed in 10% polyacrylamide, 8 M urea gel at constant power of 24 watts for 2 h and analyzed by autoradiography.
Templates for in Vitro Transcription, BOXB RNA Synthesis, and Cross-linking-For in vitro transcription assay, a DNA fragment (35,253-35,718) carrying pL-nutL was amplified from DNA/HindIII markers (Fermentas, Hanover, MD) by PCR using GeneAmp (Roche Diagnostics) and the following two primers: TCA GAT CTC TCA CCT ACC AAA C and AGG GCG GTT AAC TGG TTT TG. The PCR product was then purified using the QIAquick PCR purification kit (Qiagen). For the nutR boxB template, two complimentary DNA oligonucleotides with 5Ј EcoRI and 3Ј HindIII restriction sites (MWG Biotech) annealed. The oligonucleotides used were GGC GAA TTC ACA TTC CAG CCC TGA AAA AGG GCA TCA AAT AAG CTT GCG and CGC AAG CTT ATT TGA TGC CCT TTT TCA GGG CTG GAA TGT GAA TTC GCC. After annealing, the boxB template was cloned into pGEM-3Z (Promega). Both pGEM-3Z vector and the double-stranded template were digested with EcoRI and HindIII restriction enzymes at 37°C for 30 min. Linearized pGEM-3Z was purified with the PCR purification kit (Qiagen), and the boxB template was purified by nucleotide removal kit (Qiagen). 6 pmol of boxB template and 2 pmol of linearized pGEM-3Z were ligated in a 16°C water bath for 6 h. The ligated mixture was then transformed into XL-1B supercompetent cells (Stratagene) and grown on LB ϩ 50 g/l ampicillin plate at 37°C overnight. Colonies were purified and miniprepped, and the presence of boxB template was confirmed by sequencing. For cross-linking, DNA fragment (35,409 -35,659) was amplified with the following primers: ATA CAG ATA ACC ATC TGC GGT GAT and TGA ACG AAA ACC CCC CGC GAT TGG. 5Ј-Phosphates were removed by addition of calf intestine phosphatase (Fermentas). Template was then labeled with [␥-32 P]ATP and T4 polynucleotide kinase (New England Biolabs, Beverly, MA).
BOXB Synthesis and Gel Mobility Shift Assays-10 g of pGEM-3Z/ nutRboxB was digested with HindIII at 37°C for 30 min and then purified with the PCR purification Kit (Qiagen). In vitro transcription was performed with 6 g of the template using the T7-MEGAshortscript kit (Ambion) for 4 h. Template was eliminated by incubation with 2.5 units of RNase-free DNase I (Stratagene) at 37°C for 30 min. The reaction mixture was then purified with the nucleotide removal kit (Qiagen), loaded onto a 10% polyacrylamide, 8 M urea gel, and analyzed by autoradiography.
␣-32 P-Labeled BOXB (100 nM) was incubated for 10 min on ice with wild type or K106A/K107A, K106R/K107R, and K106D/K107D Nun (200 nM) in a 20-l reaction containing 20 nM Tris acetate (pH 7.9), 2 mM magnesium acetate, 100 mM potassium acetate, 25 mM NaCl, 0.1 mg/ml  2. A, the bacteriophage pL and pR operons. HK022 Nun and N compete for binding to NUTL and NUTR. Nun induces transcription arrest at sites distal to nut; N suppresses termination. rIII, an RNase III processing site, is located between nutL and the beginning of N (25)(26)(27). B, hypothetical structure of a Nun-arrested TEC. The Nun NH 2 -terminal ARM domain binds the BOXB sequence in the nascent transcript, facilitating the binding of the COOH terminus to RNAP and to DNA template. Contacts to RNAP involve Nun residues, His 93 , His 98 , and His 100 and require Zn 2ϩ ion. Trp 108 is shown intercalating into DNA, aided by positive charges on Lys 106 and Lys 107 . Nun binding to BOXB blocks binding of N antitermination protein.
DNA Cross-linking-500 pmol of wild-type Nun (Nun 110C) or mutant Nun (K106A/K107A110C, K106R/K107R110C, K106D/K107-D110C) in 50 mM sodium phosphate (pH 7.0) was reduced with 50 mM DTT at 37°C for 30 min and desalted with a G25 column essentially as described previously (10). The buffer was then exchanged with 0.1 M sodium borate (pH 8.4). The following procedures were carried out in the dark under red light illumination. N-[(2-Pyridyldithio)ethyl]-4azidosalicylamide (AET) (Molecular Probes) was dissolved in Me 2 SO and serially diluted with phosphate-buffered saline. To attach AET to Nun C110, wild-type or mutant Nun (20 l) was incubated with 2 nmol of AET (10 l) on ice for 30 min. The reaction product was passed through a G25 desalting column to remove excess unreacted AET. For the preparation of TEC, transcription was allowed to proceed to the ϩ15 position of the template (10 nM) by incubating RNAP (20 nM) at 32°C for 20 min with a mixture of 5 M ribonucleoside triphosphates (rNTPs), lacking rUTP in 30 l of transcription buffer containing 20 mM Tris acetate (pH 7.9), 60 mM potassium acetate, 2 mM magnesium acetate, 0.2 mg/ml bovine serum albumin, 100 M zinc chloride, and 5% glycerol. AET-Nun (500 pmol) was then added to the paused elongation complex. At this concentration, Nun binding to TEC is independent of BOXB. After 30-min incubation, the reaction mixture was UV-irradiated with a hand-held UVGL-25 lamp (UVP, San Gabriel, CA) for 2 min at the short wavelength setting with a maximum wavelength of 254 nm and stopped with 5 mM EDTA. As a control, AET-Nun was incubated with template DNA in the absence of RNAP and rNTPs and irradiated with UV light in the same manner described above. The DNA was purified with QIAquick PCR purification kit (Qiagen). For wild-type Nun, the DNA was divided into two samples, one of which was treated with 50 mM DTT at room temperature for 10 min to cleave AET from Nun. All samples were then heated to 95°C to melt template DNA and loaded onto 10% polyacrylamide, 8 M urea gel. 32 P-Labeled template DNA was loaded as marker.
Gel Filtration Chromatography-For RNAP binding, 5 M RNAP (Epicenter) was incubated with 10 M wild type, K106A/K107A, K106R/ K107R, K106D/K107D, or T-Nun (Nun V96) at 4°C for 10 min in 50 l of 50 mM sodium phosphate (pH 7.0), 100 M zinc chloride. For gel preparation, 1 g of Sephadex G-100 superfine (Amersham Biosciences) beads was swollen overnight with 30 ml of 50 mM sodium phosphate (pH 7.0), 100 M zinc chloride. For the column, 1.5 ml of Sephadex G-100 gel slurry was loaded onto a Bio-Spin Chromatography column (Bio-Rad). The excess packing buffer was initially drained by gravity and then by centrifugation at 1,000 ϫ g for 4 min. For gel filtration, the reaction binding mixture was applied to the Sephadex G-100 column and centrifuged at 1,000 ϫ g for 3 min. The eluates were electrophoresed on 15% SDS-polyacrylamide gels and analyzed by Coommassie Blue staining.

Effect of Mutations in Nun Residues Lys 106 and Lys 107 on
Termination Activity in Vivo-Nun arrests transcription on the chromosome at many different sites distal to the nut sequences (21,22). Interaction between Nun and DNA template does not, therefore, involve specific DNA base contacts. Instead, it depends on interactions between the planar aromatic Nun Trp 108 residue and double strand DNA. Trp 108 may intercalate into DNA (15). Nun residues Lys 106 and Lys 107 lie adjacent to Trp 108 (Fig. 1). We asked whether these basic amino acids have a role in Nun-dependent transcription termination in vivo, perhaps by neutralizing negative DNA charges and allowing W108 to interact with template.
Plasmids expressing Nun mutants were transformed into E. coli strains carrying Nun-sensitive transcriptional chromosomal fusions. N8499 carries pL-nutL-lacZ, and N7499 carries pR-cro-nutR-tR1-galETK. The cI857 repressor controls both promoters. Transcription from the pL or pR promoter was initiated by shifting cultures from 32 to 42°C, which inactivates the repressor. Nun termination efficiency was determined by ␤-galactosidase (Table I) or galactokinase assay (Table II).
As shown in Table I, replacement of either or both lysines with arginine, another basic amino acid (K106R, K107R, and K106R/K107R) had no effect on Nun activity in the pL-nutL operon. Alanine substitution of either Lys 106 or Lys 107 slightly decreased termination activity. Substitution of both lysines (Nun K106A/K107A) reduced transcription termination to 79% wild-type levels. Replacement of a single lysine with a negatively charged aspartate residue also reduced Nun activity; K106D and K107D were 89 and 91% as active as wild-type Nun, respectively. However, Nun K106D/K107D, which carries negative charges at both residues 106 and 107, was only 49% as active as wild-type Nun.
The effects of mutations at Lys 106 and Lys 107 on termination in the pR-nutR operon are shown in Table II. Substitution of one or both lysine residues with arginine did not dramatically reduce activity. Nun mutants K106R, K107R, and K106R/ K107R all terminated transcription with at least 90% efficiency. The effects of alanine and aspartate substitutions were more dramatic. K106A and K107A were 85 and 70% as active as wild type, respectively, and K106A/K107A had no termination activity. Substitution of either Lys 106 or Lys 107 with aspartate reduced Nun termination efficiency 3-4-fold. As reported previously (16), substitution of both lysines with aspartate increased galactokinase levels more than 2-fold rel- ative to Nun Ϫ controls. We suggest that Nun K106D/K107D is inactive as a termination factor but binds to BOXB at NUTR and sterically inhibits access of Rho to the tR1 terminator (16). Overall, these results indicate that the positive charges carried by Lys 106 and Lys 107 facilitate Nun termination activity, presumably by interacting with negatively charged DNA template. It also confirms that the pL-nutL operon is more forgiving of Nun mutations than the pR-nutR operon, a point to which we return under "Discussion." Substitutions at Nun Positions 106 and 107 Affect Exclusion-Nun residues Lys 106 and Lys 107 also play a key role in exclusion (Table III). Phage failed to form plaques on strains expressing wild-type Nun or Nun variants K106R, K107R, and K106R/K107R. Similarly, Nun K106A and K107A completely excluded . However, single aspartate substitutions (K106D, K107D) partially inhibited exclusion (efficiency of plating ϭ ϳϽ10 Ϫ4 ). Furthermore, Nun K106A/K107A and K106D/K107D failed entirely to block plating (efficiency of plating ϭ 1.0). These results indicate that position 106 or 107 must carry a positively charged amino acid for Nun to block transcription on infecting chromosomes.
Effect of Mutations on Lys 106 and Lys 107 in Transcription Arrest in Vitro-We next tested the Nun mutants in a minimal in vitro transcription system. In this assay, Nun arrests transcription without releasing stalled TEC. As template, we used a 466-bp (35,253-35,718) DNA fragment containg pL-nutL, which was amplified from genomic DNA/HindIII (Fig. 3). Nun arrest activity is indicated by reduction in run-off transcripts (RO) and the appearance of specific short RNAs (NA; Fig. 3,  lanes 3, 4, and 6). As reported previously (9), T-Nun (T), a truncated form of Nun that lacks the COOH-terminal 13 amino acid residues, was entirely inactive (Fig. 3, lane 2). Nun K106A/ K107A (AA) and K106R/K107R (RR) both arrested transcription, although the former was less active than K106R/K107R (Fig. 3, lanes 4 and 6). Nun K106D/K107D failed completely to arrest transcription (Fig. 3, lane 5). This in vitro data further supports the hypothesis that the electrostatic interaction between Nun residues Lys 106 and Lys 107 and the DNA template facilitates Nun-dependent transcription arrest.
BOXB and RNAP Binding by Wild-type and Nun Lys 106/107 Mutants-In addition to contacting DNA template, Nun transcription termination entails two additional binding reactions. First, Nun is recruited to the TEC by binding of the NH 2terminal ARM domain to the BOXB elements of the nascent transcript NUT sequences (Fig. 2B; Refs. 7 and 8). Second, Nun binds to RNA polymerase, either in the TEC or in solution (9, 10). Binding is Zn 2ϩ -dependent and is promoted by Nun resi-dues His 93 , His 98 , and His 100 (Fig. 2B). 2 We asked whether Lys 106 and Lys 107 played a role in these interactions. The ability of Nun mutants, K106A/K107A, K106R/K107R, and K106D/K107D to bind BOXB was tested by a native gel mobility shift assay (Fig. 4A). All mutants bound BOXB with approximately wild-type efficiency. Averaging three independent experiments, under these conditions wildtype Nun bound 47% (Ϯ6.5%) of input BOXB, whereas Nun K106D/K107D bound 48% (Ϯ15.7%).
Binding of wild-type and mutant Nun to RNAP was compared using size exclusion chromatography on a Sephadex G-100 superfine column (Fig. 4B). Under our assay conditions, wild-type Nun binds to RNAP (Fig. 4B, lane 2), whereas T-Nun is retained in the column (Fig. 4B, lane 6). Wild-type Nun, the arrest-proficient Nun mutants, K106A/K107A and K106R/ K107R, and the arrest-defective Nun mutant K106D/K107D, all bound RNAP (Fig. 4B, lanes 2-5). These results indicate that  substitutions at Lys 106 and Lys 107 that abrogate transcription arrest do not affect the ability of Nun to bind to BOXB or RNAP.
Lys 106 and Lys 107 Promote Nun-DNA Contacts-Nun can be photochemically cross-linked to DNA template in a Nunarrested TEC (10). Cross-linking is dependent on Nun residue Trp 108 . The results reported above suggest that Lys 106 and Lys 107 help position Trp 108 at the template by neutralizing DNA negative charges.
To demonstrate cross-linking, we used Nun derivatives carrying an additional C110 residue linked by a disulfide bond to the photoreactive cross-linker, AET ( Fig. 5; Ref. 10). The 251 bp pL-nutL DNA fragment described under "Materials and Methods" was used as a transcription template after dephosphorylation and 5Ј-end labeling with [␥-32 P]ATP. Transcription was performed as described in the legend to Fig. 5. As reported previously (10), AET-Nun did not cross-link [ 32 P]DNA template in the absence of a TEC (Fig. 5, lane 2). Incubation of the TEC with Nun110C yielded an Nun-AET-[ 32 P]DNA complex that migrated slower in the gel than the labeled template marker (Fig. 5,  lane 3). As expected, the complex was sensitive to DTT, which reduces the disulfide bond in Nun-S-S-AET-[ 32 P]DNA complex (Fig. 5, lane 4). Nun K106R/K107R110C showed the same crosslinking pattern as wild-type Nun (Fig. 5, lanes 3 and 7). K106A/ K107A110C formed less complex with template DNA than wildtype Nun (Fig. 5, lanes 3 and 5). Finally, K106D/K107D110C failed to cross-link template (Fig. 5, lane 6). We conclude that the transcription termination/arrest defect of K106D/K107D is explained by its inability to contact template DNA. DISCUSSION Prophage HK022 Nun protein and phage N protein bind the same NUT sequences in nascent transcript (7,21). However, Nun induces transcription arrest, whereas N both accelerates the rate of transcription elongation and suppresses transcription termination (6). The functional differences between Nun and N are specified by their unique carboxyl-terminal regions (23). The COOH terminus of Nun is required for Nun transcription arrest both in vivo and in vitro. A truncated variant of Nun, T-Nun, which lacks Nun residues 97-109, is completely inactive in blocking transcription elongation (9). T-Nun binds BOXB sequences within NUTL and NUTR (9) but, unlike wildtype Nun, fails to interact with RNAP (9, 10) or DNA template. 3 The contribution of Nun carboxyl-terminal residues to transcription arrest has been explored by genetic and biochemical analyses. Histidine residues His 93 , His 98 , and His 100 coordinate Zn 2ϩ (15) and are required for binding to RNAP. 2 The penultimate Nun residue, Trp 108 , plays a critical role in Nun arrest. There are suggestions that Trp 108 may intercalate in the template, allowing Nun to act as a brake to TEC translocation. Thus Nun does not arrest transcription on single strand DNA (15). Furthermore, although Nun W108A or Nun W108L fails to terminate transcription, substitution of Trp 108 with another aromatic planar amino acid, tyrosine, is innocuous (15). Finally, Nun arrests transcription at many sites promoterdistal to NUT (21,22). This suggests that specific amino acidnucleotide interactions are not involved in Nun transcription 3   arrest. Nun W108A, although it interacts with a TEC, does not form stable contacts with DNA (10). Interaction of Nun with template occurs only in the context of a TEC (Ref. 10 and this work) and entails contacts between Nun and RNAP. Nun H93A, like W108A, does not cross-link template (10).
Two basic residues, Lys 106 and Lys 107 lie adjacent to Trp 108 (Fig. 1). We present evidence above that Lys 106 and Lys 107 facilitate the proper positioning of the Nun carboxyl terminus near the template by electrostatic interactions with DNA. Thus, K106D/K107D is defective for exclusion and for transcription termination in vivo and fails to arrest transcription in vitro. Importantly, K106D/K107D does not cross-link with DNA template. In contrast, K106R/K107R, which preserves the basic charges at positions 106 and 107, is fully active in transcription arrest and in cross-linking DNA. K106A/K107A crosslinks template DNA with an efficiency intermediate between wild-type or K106R/K107R and K106D/K107D.
The defect in the Nun pathway of K106D/K107D is accounted for entirely by inability to interact with template DNA. The mutant protein binds BOXB and RNAP in vitro with wildtype efficiency. Evidence for binding of K106D/K107D to BOXB in vivo has been reported previously (16). We conclude that Lys 106 and Lys 107 act to neutralize negative charges on DNA template, allowing Nun to interact with template and to block translocation of the TEC.
In general, Nun proteins mutated at residues 106 and 107 are less active at nutR than at nutL. Thus, K106D/K107D and K106A/K107A induce 49% and 79% termination at nutL, respectively, but are entirely inactive at nutR. In this respect, Nun mutants act like host nus and rpoC mutants, which suppress Nun at nutR but have little effect at nutL (2,24). Two factors contribute to the difference between nutL and nutR. First, nutL lies only 34 bp from pL, whereas nutR is separated by 227 bp from pR (14). The proximity of nutL to its cognate promoter may block access of host transcription repair-coupling factor to the Nun-arrested TEC, thus prolonging the half-life of the complex (14). Second, the pL operon includes the RNase III-sensitive site, rIII, which lies just distal to nutL ( Fig. 2A; Refs. [25][26][27]. rIII enhances Nun activity, perhaps by making contacts with Nun residues. 4 Alanine scanning of the Nun 13 carboxyl-terminal amino acids (Asn 97 -Ser 109 ; Fig. 1) revealed that an R102A substitution reduced termination efficiency to 67% at pL-nutL and completely abolished activity at pR-nutR (data not shown). In contrast to Lys 106 and Lys 107 , however, substitution with another basic amino acid (R102K) was equally disabling. The role of Arg 102 in transcription termination, therefore, appears to be unrelated to its charge.