Module 4.5.3.1

Nus Factors of Escherichia coli

RANJAN SEN,* JISHA CHALISSERY, AND GHAZALA MUTEEB

[SECTION EDITOR: IRINA ARTSIMOVITCH]

Posted January 18, 2008

Laboratory of Transcription Biology, Center for DNA Fingerprinting and Diagnostics, ECIL Road, Nacharam, Hyderabad 500076, India

*Corresponding author. Mailing address: Laboratory of Transcription Biology, Center for DNA Fingerprinting and Diagnostics, ECIL Road, Nacharam, Hyderabad 500076, India. Phone: (91) 40 27151344, Fax: (91) 40 27155610, E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it .

Nus factors are a set of well-conserved proteins in bacteria, involved in transcription elongation, termination, antitermination, and translation. The four Nus factors, NusA, NusB, NusE, and NusG, are encoded by the genes nusA, nusB, nusE (rpsJ), and nusG, respectively. Originally, Escherichia coli host mutations involved in bacteriophage λ N-mediated antitermination were mapped to the nusA (nusA1), nusB (nusB5 and nusB101, etc.), and nusE (nusE71) genes, and hence, these genes were named nus for N utilization substances (36, 37, 39, 59). Among these three genes, nusA and nusB encode unidentified factors, whereas nusE is known to encode a ribosomal protein, S10. The involvement of nusG in N-mediated antitermination via a suppressor (nusG4) of one of the nusB mutations (in nusB101) was identified (103) and later confirmed by NusG depletion assays (123). Subsequently, the Nus factors were purified by different groups (48, 50, 63, 66, 98, 105), and their roles in different host functions, such as ribosomal operon transcription, Rho-dependent and -independent transcription termination, and translation, were elucidated (29, 100). The essentiality of Nus factors for the viability of E. coli has been envisioned from their highly conserved nature among different bacteria and their involvement in various essential cellular functions (32, 85, 122). By systematic gene replacement methods, it was shown that, except NusB, which is conditionally lethal, all the Nus factors are essential for the viability of E. coli (20).

The locations of the Nus genes on the chromosome reveal that these genes belong to four different operons (Fig. 1). nusG and nusE are clustered with ribosomal protein genes, and it is likely that this arrangement is required for transcription-translation coupling, a process in which the ribosome loads onto the nascent RNA as soon as its attachment site is accessible and translates the RNA at a rate that is synchronous with the rate of transcript elongation in E. coli. The nusA gene belongs to the metY-infB operon and may be involved in the regulation of the adjacent rbfA gene, which codes for a ribosome binding factor (26). The significance of the clustering of the nusB gene with the riboflavin biosynthesis (rib) genes is not clear.

Fig. 1Operonic arrangements of all four nus genes in the E. coli genome. Positions of the respective nus genes are boxed. The directions of the arrows indicate the directions of transcription. Numbers at both sides indicate the nucleotide sequence positions.

A summary of the different biochemical properties, various functions, and interacting partners of the Nus factors is shown in Table 1.

Table 1Summary of different properties of the Nus factors.

BIOCHEMISTRY AND STRUCTURES OF Nus PROTEINS

NusA

NusA is a 55-kDa protein that is capable of binding to (i) E. coli core RNA polymerase (RNAP) (46), (ii) the transcription antitermination protein N (47), (iii) specific RNA sequences in the nut site of λ RNA (38, 82), and (iv) the ribosomal operon leader region RNA (4). This multiple-binding mode of NusA is reflected in its multidomain structural organization (Fig. 2A). This 495-amino-acid protein has three RNA binding motifs, S1, K homology 1 (KH1), and KH2; two C-terminal acidic regions, acidic region 1 (AR1) and AR2; and an N-terminal RNAP binding domain (17, 41, 71, 117). Except the C-terminal AR1 and AR2 domains, the domains are highly conserved among the NusA proteins from different species. A protein BLAST search revealed that these C-terminal domains (~160 residues) present in E. coli NusA and the NusA proteins of other enterobacteria are missing in several other species. These two domains actually block the binding of free E. coli NusA to RNA, and interactions between AR2 and the α subunit (α C-terminal domain) of RNAP in the elongation complex (EC) activate the RNA binding property of the S1 and KH domains (69, 71). NusA from Mycobacterium tuberculosis can bind to RNA on its own, as it does not have these C-terminal domains (14). The AR1 domain was also found to interact with the central part of λ N protein, but this interaction failed to eliminate the inhibitory effect on the RNA binding (15). The three RNA binding domains, S1, KH1, and KH2, form a continuous, positively charged surface suitable for binding to RNA. Although several NusA-specific mutations in the β and β' subunits of RNAP have been obtained, the binding surface has not been defined conclusively (56, 57). Chemical cleavage experiments (107) suggest that the surface for σ⁷⁰ interaction on the core RNAP overlaps with that for NusA interaction, which is consistent with the model of direct replacement of the σ subunit by NusA (42) at the onset of transcription elongation.

Fig. 2Crystal structures of NusA domains and complexes with RNA and N protein. (A) Cartoon showing different functional domains of E. coli NusA. The NTD and the Src homology (S1), KH1, and KH2 domains are highly conserved among different NusA molecules. Domains AR1 and AR2 are found in a few species, including E. coli. Numbers indicate amino acid positions. (B) The crystal structure of conserved domains of NusA from T. maritima (Protein Data Bank [PDB] no. 1L2F) (97) shows that each functional domain has distinct tertiary folds and corresponds to a separate structural domain, as indicated. (C) Two separate nuclear magnetic resonance structures of AR1 and AR2 domains (PDB nos. 1WCL and 1WCN, respectively) (33) of E. coli. (D) Crystal structure of M. tuberculosis NusA (pink) with the stem-loop RNA (black) from boxC of an M. tuberculosis rRNA operon (corresponding to PDB no. 2ASB) (14). The structure revealed that both the RNA binding domains KH1 and KH2 take part in the interaction with the RNA and that the interaction destabilizes the stem-loop structure of boxC. (E) Solution structure of the central part of N protein (shown as a green thread) from bacteriophage λ in a complex with the dimeric interface of the AR1 domain of E. coli NusA (PDB no. 1U9L) (15). The dimeric interface was found to be formed from two AR1 domains coming from two NusA molecules. The region encompassing residues 352 to 419 of one NusA molecule (in orange) dimerizes with the region encompassing residues 352 to 421 of the second NusA molecule (blue). This indicates that in the antitermination complex with λ N protein, NusA may form a dimer. Amino acid numbers of the relevant region of the λ N protein are indicated.

The high-resolution crystal structures of NusA proteins obtained from Thermotoga maritima (97, 117) and M. tuberculosis (44) revealed that NusA is a rod-shaped, elongated molecule consisting of four distinct structural domains (the N-terminal domain [NTD], S1, KH1, and KH2) corresponding to its functional domains (Fig. 2B). The three RNA binding domains (S1, KH1, and KH2) are held by the interdomain interactions as a rigid body and are connected to the NTD via a linker so that the orientation of the RNA binding domains can be dictated by the interaction of the NTD with the EC. A five-stranded β-barrel (S1 domain) and βααβ motifs of the two KH domains of NusA are widespread among different nucleic acid binding proteins that form a positively charged surface to accommodate nucleic acids (44, 117). The crystal structure of the M. tuberculosis NusA-nut RNA complex revealed (Fig. 2D) that both the KH domains bind to the RNA and destabilize the secondary structures in the nut site (14). The structures of the C-terminal AR1 and AR2 domains of E. coli NusA, both in free form (Fig. 2C) (33) and in complex with λ N (Fig. 2E) (15), are available; these domains are each composed of five helices, and there are no significant conformational changes between the bound and free forms (Fig. 2C and D). As these domains interact with different proteins (e.g., RNAP and λ N), their role has been inferred to be that of a simple scaffold to bring the two interacting proteins into proximity.

NusB

E. coli NusB, a ~16-kDa monomer, is highly abundant in the cell (106). This factor is an important component of the ribosomal and λ N antitermination complex (40, 79, 99, 100). In the antitermination complex, the N-terminal basic region of NusB most likely binds to a 12-nucleotide RNA sequence corresponding to the conserved element called boxA, which is present in the ribosomal operon or in the nut site of the DNA from lambdoid phages (79). The binding affinity and preference of NusB for the boxA RNA sequences are enhanced by ~100-fold when NusB forms a complex with NusE (76). The recognition of the boxA sequence by NusB plays the most important role in recruiting other Nus factors to the antitermination complex.

Two solution structures of E. coli NusB have been reported (2, 54). Both of these structures revealed that NusB is an all α-helical protein, but they differ in the tertiary folds. The earlier solution structure (Fig. 3A) (2) shows a stronger resemblance to the crystal structures of NusB proteins from M. tuberculosis (43) and T. maritima (16), which comprise two globular N- and C-terminal subdomains. Interestingly, NusB has a very unique structure, with no similarity to other known structures, and it does not contain any known RNA binding structural motif. It is possible that the N-terminal arginine-lysine-rich RNA binding region of NusB adopts more favorable conformational changes upon interactions with NusE and RNA, which are not apparent in the free-NusB structure. The exact mode of NusB-RNA interactions requires the resolution of the structure of the NusB-RNA-NusE complex.

Fig. 3Structures of NusB and NusE. (A) Solution structure of E. coli NusB (PDB no. 1EY1) (2), consisting of N- and C-terminal globular domains. This solution structure shows greater resemblance to the crystal structures of homologous NusB proteins from M. tuberculosis (43) and T. maritima (16). (B) Structure of NusE (S10) as part of a 30S ribosome particle (chain J of the coordinate available from PDB no. 1J5E) (19). The RNP fold of NusE may also be important for interacting with RNAP in the transcription EC.

NusE

NusE is a ribosomal small-subunit protein, also called S10, and is one of the first proteins found to be a component of both the ribosome and the transcription antitermination complex (80, 100). This 12-kDa protein exists as an unstructured monomer in the free form, which limits the structural studies of this protein in isolation (45, 50, 70). It is possible that it forms a specific structure in the presence of other proteins in the antitermination complex, as observed when it is a part of the ribosome (see below). It forms a stable heterodimer with NusB and enhances the affinity of this complex for the boxA sequence on the RNA. The NusB-NusE-boxA complex serves as a platform for the assembly of other Nus factors into the antitermination complex (50, 70, 79).

As a component of the ribosome, S10 (NusE), together with S3 and S14, forms a tight cluster at the back of the head of the 30S subunit. The structure of the ribosome reveals that the S10 protein inside the ribosome is composed of a globular domain, which is located on the surface of the 30S subunit, and a long extended β-hairpin structure that penetrates into the inner core (Fig. 3B). S10 participates in extensive interactions with 16S RNA by involving about 25% of its surface area in this purpose (19).

NusG

NusG is a 21-kDa monomeric protein involved in transcription elongation and termination and interacts with RNAP and transcription termination factor Rho (88). The homology model of E. coli NusG derived from the crystal structure of Aquifex aeolicus NusG reveals that it comprises two distinct globular domains connected by a flexible linker (Fig. 4A) (61, 101). The N-terminal ribonucleoprotein (RNP)-like domain (domain I) is proposed to be involved in protein and nucleic acid binding. A. aeolicus NusG has a unique ~70-residue-long positively charged immunoglobulin-like central domain which is replaced by a loop known as the appended minidomain in E. coli NusG (61, 101). This domain is believed to serve as a structural element rather than being involved in residue-specific contact with other proteins (90). The KOW motif, which was first discovered in the C-terminal domain of NusG (domain II), is predicted to interact with both RNA and proteins concomitantly since different areas of this domain have been implicated in these interactions (61, 101). The flexible linker between these two domains may allow independent degrees of freedom for each of the N- and C-terminal domains, which enables interdomain cross talk in response to the interactions with external factors. The 27-residue-long KOW motif of NusG shows homology to the different ribosomal proteins RL24 (from bacteria and chloroplasts) and RL26/RL27 (from eukaryotes) and to eukaryotic proteins Spt5 and DRB sensitivity-inducing factor (64, 101, 110).

Fig. 4Crystal structures of NusG and its paralog RfaH. (A) Homology model of E. coli NusG (101), which is characterized by two globular domains, I and II, connected by a flexible linker. This linker most likely ensures degrees of freedom between the two domains and their respective orientations in the space. (B) Crystal structure of RfaH (PDB no. 2OUG) (11). The sequence of N-terminal domain I of RfaH is similar to that of domain I of NusG, and also the domain folds in a similar way as NusG domain I. The C-terminal domain II of RfaH is very different and folds into an α-helical coiled-coil structure, in contrast to the β-barrel structure of NusG, and comes over the domain I in the tertiary structure. Unlike domain II in NusG, in this conformation of RfaH, domain II blocks the access to domain I. (C) RfaH domain I is modeled onto the coiled-coil domain of the β subunit of RNAP, where it is shown that domain I is now unmasked from domain II of RfaH upon interaction with the EC. Different structural elements are indicated.

NusG homologs from various organisms show affinity for different types of nucleic acids. A. aeolicus NusG and T. maritima NusG bind to double-stranded DNA (dsDNA), single-stranded DNA, and rRNA, with T. maritima NusG showing strong preference for dsDNA. On the other hand, E. coli NusG binds to dsDNA and rRNA but not to single-stranded DNA (61, 83, 86). Although in solution NusG is present as a monomer (101), domain-swapped dimeric and polymeric forms have been observed in some A. aeolicus NusG crystals (61). The surfaces for the interaction of NusG with its partners, like RNAP and Rho, are not defined, and neither has the role of each of the domains in transcription elongation and termination been worked out. However, several point mutations in domain II that make it defective for Nun-mediated (23) and Rho-dependent termination (52) have been isolated.

RfaH, the NusG paralog, has two domains, with the N-terminal domain showing much similarity to domain I of NusG while the C terminus of RfaH is folded into an α-helical coiled-coil domain instead of the β-barrel structure of NusG (Fig. 4B) (11). Based on the recent structure of RfaH modeled on the EC (Fig. 4C) (11), it can be hypothesized that during transcription elongation, NusG interacts with various transcription factors via domain II while being anchored to the β subunit coiled-coil domain of the EC via domain I.

FUNCTIONS OF Nus PROTEINS

Transcription Elongation

NusA and NusG are the integral part of the transcription EC that regulates the rate of elongation, and therefore, these proteins function as general transcription elongation factors in all the operons (89). NusA reduces the rate of elongation by enhancing the pausing of the EC at specific sequences, whereas NusG increases the rate of elongation by reducing the pausing (22, 24).

During elongation, RNAP pauses at different sequences. Pausing sequences are not very well defined, but pause sites can be classified broadly into two classes; at class I pause sites (like the his pause site), the nascent RNA forms a hairpin near the RNA exit channel and interacts with the EC, and at class II pause sites (the ops and U-tract sites), due to the weak RNA-DNA hybrid sequences, the EC tends to backtrack (5). NusA was observed to enhance the class I pauses, whereas NusG was found to be more effective against class II pauses than class I pauses (5). These two contrasting behaviors can be explained by the differential modes of interaction of these factors with the EC. Upon the binding of NusA to RNAP through its N-terminal domain, the RNA binding domains of NusA become accessible to interact with the 5' end of the nascent RNA. The interaction of the C-terminal AR1 and AR2 domains with the α subunit of RNAP may also help the RNA binding domain of NusA to interact with the RNA (69). The interaction of NusA with the single-stranded region in the upstream end of the nascent RNA allows a stable hairpin to form near the RNA exit channel by preventing unwanted pairing with the 5' end of the RNA. The destabilization of the RNA secondary structure at the NusA-interacting sites on the RNA has been proposed based on the crystal structure of the NusA-RNA complex (Fig. 2D) (14). On the other hand, the likely interaction of NusG with RNAP near the active center helps the RNAP to hold the 3' end of the nascent RNA at weak RNA-DNA hybrid sequences and thereby prevent the backtracking of EC at the class II pause sites.

The bacterial transcription elongation factor RfaH suppresses the polarity of polycistronic operons involved in the expression of virulence factors (7). RfaH is recruited to the EC at the operon polarity suppressor (ops) site, a12-bp conserved sequence, through sequence-specific contact with the nontemplate DNA strand and was proposed to interact with the coiled-coil region of the β' subunit, which is about 75 Å away from the active center of RNAP (11). The RfaH-modified EC shows enhanced rate of elongation by suppressing pausing and reduced termination at Rho-dependent and Rho-independent terminators (6, 8). RfaH also inhibits the pyrophosphorolysis and increases the affinity of nucleoside triphosphate binding during elongation (104).

Hairpin-Dependent Transcription Termination

The canonical signal for intrinsic or hairpin-dependent termination at the end of the operons is characterized by a GC-rich palindromic element followed by an oligo(T) sequence. In the nascent RNA, this sequence forms an RNA hairpin at the RNA exit channel, followed by seven to nine nucleotide U residues situated in the RNA-DNA hybrid (81). It is believed that the oligo(dT) tract induces the pausing of the EC and allows the hairpin to form at the RNA exit channel. The formation of the RNA hairpin at this site either inactivates the EC (51) or directly induces melting at the 5' end of the weak RNA-DNA hybrid (62). NusA improves the termination efficiency at various intrinsic terminators (13, 34, 93, 111). Because NusA decreases the overall elongation rate, it is believed that it prolongs the pausing half-life at the terminator and thereby improves the termination efficiency (34, 93). It has also been proposed that NusA stabilizes pause hairpin-flap interaction, which by widening the RNA exit channel may allosterically affect RNAP’s active site (108). It can be postulated that in the NusA-pause hairpin-EC complex, NusA may interact with the flap domain of the β subunit.

More recently, from the crystal structure of the NusA-RNA complex (Fig. 2D) (14) and the results of cross-linking experiments with the EC at the terminator site (51), it has been proposed that NusA, upon interaction with single-stranded RNA at the 5' end behind the moving EC, chaperones the terminator hairpin to fold properly in the RNA exit channel by preventing the unwanted pairing with the upstream part and nonspecific interactions with the nearby surface of the EC. This mode of action of NusA during termination very well explains the NusA dependence of some of the terminators with relatively unstable stem structures (the presence of bulges in the stem) (87).

Ribosomal protein L4 in conjunction with NusA can induce termination at the attenuator position of the leader region of the S10 operon comprising several stem-loop structures (94, 95, 120, 121), the mechanism of which is yet to be explored.

Rho-Dependent Transcription Termination

Rho is a homohexameric RNA-DNA helicase or translocase that dissociates RNAP from the DNA template and releases the RNA. The RNA-dependent ATPase activity of Rho provides free energy for these activities (9, 29, 88). Rho-dependent termination involves a series of sequential events. At first, Rho binds to the C-rich regions of the nascent RNA, called rut (Rho utilization) sites, through its primary RNA binding domain (10, 73). This binding leads to the positioning of the RNA into the secondary RNA binding domain, which in turn activates the ATPase activity. Rho is capable of translocase and RNA-DNA helicase functions using the free energy derived from ATP hydrolysis (18), eventually releasing the RNA on which Rho is loaded. It is commonly believed that the translocase-helicase activity of Rho is instrumental in pulling out RNA from the elongating RNAP (88).

Rho can work efficiently in a purified in vitro system; however, NusG enhances the termination efficiency and advances the point of terminations at the promoter-proximal sites, which are otherwise not preferred by Rho in the absence of NusG (67). The involvement of NusG in Rho-dependent termination was first identified genetically in the strains which either overexpress NusG or are depleted of NusG (102, 103). The retardation of Rho by a NusG column (67) suggested physical contact between Rho and NusG, and the interaction was further quantified later to have an equilibrium constant of ~10⁻⁸ M⁻¹ (83).

NusG causes early termination at several terminators, including the trp terminator, and can overcome the defects of certain Rho mutants (21). NusG does not improve either the rate of unwinding of the RNA-DNA hybrid or ATP hydrolysis or the RNA binding properties of Rho (78). Rho-dependent RNA release kinetics suggest that NusG is involved only in the RNA release step(s) once Rho reaches the vicinity of the EC (21, 27). The early termination that is usually observed in the presence of NusG is the consequence of this enhanced rate of RNA release. NusG becomes mandatory under in vivo conditions in which the RNA release has to be performed within a small window of time.

NusA may bind to Rho (92), although the physiological relevance of this binding is not understood. The functional involvement of NusA in Rho-dependent termination was first observed as the NusA-independent growth of E. coli strains which were defective for Rho-dependent termination (122). However, in vitro transcription assays in the presence of NusA revealed that NusA either inhibits (58) or delays (22) the Rho-dependent termination process. It is possible that NusA-mediated sequence-specific modulation of the properties of the EC affects the Rho-dependent termination at different termination points.

Transcription Antitermination in Lambdoid Phages

The antiterminator protein N from lambdoid phages, upon interacting with the nut site of the nascent RNA (28, 65) and in the presence of Nus factors, modifies the elongating RNAP in such a way to express the downstream early genes of the phage (30, 49, 74) by overcoming both Rho-independent and Rho-dependent terminators (35, 114) (Fig. 5A). The nut site contains two elements, boxA and boxB. N recognizes the GNRA tetraloop of the boxB hairpin via its arginine-rich motif at the N terminus (75). The central part of N is anchored onto the AR1 motif of NusA, whereas the C terminus interacts with RNAP (31, 71, 75). N on its own can bring about the antitermination (51, 87), but for the processive antitermination to occur in vivo, stable and cooperative interactions between N and all the Nus factors are essential (Fig. 5B). The assembly of different components of the antitermination complex may proceed through the following sequential steps (72, 109, 124). Upon the release of σ⁷⁰, elongation factors like NusA and NusG and probably NusE become part of the EC. NusE stabilizes the binding of NusB onto the boxA sequence, whereas N recognizes the boxB hairpin, after which it interacts with NusA and its C terminus gets chaperoned into the EC. The interaction of NusB with boxA may stabilize the short hairpin structure in the adjacent boxB sequence. The presence of two NusA molecules in this antitermination complex has been proposed (51), where N modifies one NusA molecule to interact firmly with the upstream shoulder of the unfolded terminator hairpin and RNAP modifies the other NusA molecule to interact with the nut RNA. This configuration impairs the terminator hairpin formation and enables the EC to overcome the termination signal. The solution structure of the NusA-N complex also revealed that the central part of N binds to the dimeric interface of two AR1 domains of NusA proteins (Fig. 2D). Therefore, it is quite possible that NusA exists as a dimer in the antitermination complex.

Fig. 5Antitermination complex of λ N protein. (A) Gene organization of early transcription region of lambdoid phages. Transcripts originating from P_L (left) and P_R (right) promoters get prematurely terminated at T_L (left) and T_R (right) terminators in the absence of N protein. N protein together with a battery of Nus factors recognizes the boxA and boxB elements of the RNA nut sites. The nature and sequences of these conserved elements from different lambdoid phages are depicted. Boxes (stem) and arrows (loop) indicate nucleotides corresponding to stem-loop structures. (B) Cartoon showing the assembly of an N antitermination complex on the nut site. Interactions of the NusB-NusE heterodimer with boxA, the N terminus of N protein with boxB, and N protein with the AR1 region of NusA have been experimentally verified. Other interaction sites in the complex have not been proven experimentally.

Source: Fig. 5A is adapted from MolecularMicrobiology (77) with permission of the publisher.

The in vivo requirement for Nus factors in Q-mediated antitermination has not been demonstrated. However, NusA stimulates Q-mediated antitermination in vitro (81). This stimulation occurs most likely by stabilizing the interaction of Q at the β subunit flap domain, and the presence of a NusA-Q complex near the RNA exit channel forms a shield over the exiting mRNA, which occludes the terminator hairpin from forming in this channel (96).

An interesting variant of N protein is Nun protein from coliphage HK022, which also interacts with the boxB hairpin via its N-terminal arginine-rich motif (28, 91). Its C terminus interacts with the EC in a different way so that, in the presence of all the Nus factors, it does not suppress termination; rather, it induces the arrest of the EC by blocking the forward translocation (55, 60, 113), and in vivo, the arrested RNAP possibly is released from the site by the action of Mfd protein, which is usually recruited at the sites of arrested ECs (112).

Transcription Antitermination at Ribosomal Operons

The transcription of rRNA operons constitutes the majority of the RNAP activity in vivo, and the growth rate of the cell depends on the efficiency of this activity. To meet the huge need for the components of the ribosome, very efficient transcription initiation and elongation strategies for rRNA transcription have been evolved. As the rRNA operons code for large untranslated regions and the RNA sequence in the spacer and leader regions do not fold into strong secondary structures (e.g., boxA and boxC sequences) (Fig. 6A), the transcription process itself is a target of premature Rho-dependent termination. To overcome these premature termination events, an antitermination system similar to that of lambdoid phages, involving different Nus factors, boxA and boxB sequences, and some ribosomal proteins, is in existence in bacteria (84, 100). However, this antitermination system is more efficient at suppressing Rho-dependent termination than intrinsic termination (1).

Fig. 6Antitermination complex at the rRNA operon. (A) Typical organization of an rRNA operon corresponding to 16S, 23S, and 5S RNAs. The antitermination signals (AT) present in the leader and spacer regions are depicted. The major elements (boxB-boxA-boxC) recognized by the different components of the antitermination complex are shown. (B) Cartoon showing the assembly of different Nus factors and ribosomal proteins on the boxA element of rRNA. Also shown is the transcription terminator Rho in the context of this antitermination complex. This complex helps the RNAP to overcome the Rho-dependent RNA release by increasing the elongation speed, by blocking the access of Rho, or by using both these mechanisms.

All the Nus factors (99, 119), together with different ribosomal proteins such as S2, S4, L1, L3, L4, and L13, are part of this antitermination system (98, 106) (Fig. 6B). These proteins assemble on the antitermination signals characterized by boxB-boxA-boxC sequences present just downstream of the rrnP2 promoter and in the spacer region between 16S and 23S RNA genes (Fig. 6A). The sequences of boxA and boxC are quite conserved, whereas boxB is a stem-loop structure with no obvious sequence conservations. Among these three elements, boxA is the most crucial and a prominent feature of most of the rRNA operons in the bacterial world (1, 12, 68, 99). The NusB-NusE heterodimer binds to the boxA sequence with high affinity (50, 79) and most likely nucleates the entire antitermination complex, whereas the elongation factors NusA, NusG, and NusE directly interact with the core RNAP (46, 66, 72). The presence of Nus factors alone may increase the elongation rate through the rRNA operon, but the addition of S-100 extract containing different ribosomal proteins is required to overcome the Rho-dependent termination (98). Recently, it has been shown that the addition of ribosomal protein S4 alone to the Nus factors produces efficient in vitro antitermination and that the efficiency improves further upon the addition of ribosomal proteins L3, L4, and L13 (106). As this efficiency still does not reach the efficiency that is obtained in the presence of whole S-100 extract, it is believed that other factors are also involved.

Although the mechanistic details of this antitermination process have not been worked out, the following interesting features of this system should be noted. This antitermination complex increases the elongation rate by more than twofold and is found to be efficient only with Rho-dependent terminators (1), and the modified EC terminates efficiently at the two strong intrinsic terminators, T1 and T2, at the end of the operon. These findings suggest that unlike the N antitermination complex, this complex has no effect on the folding of terminator hairpins. It can be speculated that the enhanced elongation speed of this modified EC is a key factor for overcoming the "Rho-chase." It is also likely that this huge antitermination complex blocks the translocation of Rho along the RNA and prevents the access of Rho to the EC. It will be interesting to know how important is the assembly of this antitermination complex in the strains with defective Rho or in the bacteria, like Bacillus subtilis, where Rho is not essential.

Contribution to Translation Machinery

S10 (NusE), an integral part of the 30S ribosomal subunit, is found at the top of the subunit and interacts with the 16S rRNA through the RNP fold in its N-terminal globular domain. The central part of NusE folds into a β-hairpin loop which penetrates towards the aminoacyl-tRNA binding site and comes close to the primary tetracycline binding site of the 30S ribosome (19). Recently, a Tet^r phenotype of the bacterium Neisseria gonorrhoeae has been mapped to a point mutation in S10 leading to reduced binding affinity for the translational inhibitor tetracycline (53).

The involvement of the N antitermination complex in the translational repression of the N gene has been reported previously (115, 116). It is hypothesized that 5' end of the nut RNA may remain tethered to the antitermination complex, which blocks the ribosome loading onto the RNA.

FUTURE PERSPECTIVE

Over the last three or four decades, extensive genetic, biochemical, and structural studies of Nus factors have established the biochemical and structural properties of these factors and their functional involvement in two major cellular processes, namely, transcription and translation. Recent genomics approaches (3, 25, 118) have predicted that these Nus factors are part of a very intricate cellular interaction network and are involved in interacting with a plethora of cellular proteins, the majority of which have unknown functions. Consistent with these predictions, the operonic organizations, especially for nusA and nusB, with genes for apparently nonrelated proteins are also suggestive of diverse interaction patterns of these Nus factors. The presence of multiple functional domains with specific structural folds in NusA and NusG and the capability of these factors to interact with proteins, DNA, and RNA strongly indicate that they have the potential to interact with varied partners. On the other hand, the apparently unstructured nature of NusE in free form should enable it to fold into different conformations upon interaction with proteins and nucleic acids of diverse natures. Some genetic evidence has also implicated NusB in the translation of secretory proteins (100). Therefore, it has now become important to explore the other functions and the colocalization of Nus factors with different cellular apparatuses by using whole-genome approaches like microarrays and chromatin immunoprecipitation techniques.

Although the structural information for individual Nus factors is available, our understanding of the surfaces of these factors for interaction with RNAP and other associated factors in the antitermination and termination complexes is very poor. This information is a prerequisite to understanding the mechanisms and the role of each of the Nus factors in the complex. The essentiality of each of these factors for bacterial growth limits the application of genetic techniques to define the interaction surfaces. In the absence of crystal structures of these complexes, systematic biochemical approaches such as footprinting and site-specific cross-linking techniques will be useful to resolve these questions.

Except for RfaH, Nus factors have not yet been found to be involved in the virulence functions of pathogenic bacteria. It will be worth searching for other Nus orthologs or paralogs encoded in the prokaryotic genomes which are associated with pathogenicity. On the other hand, the functional essentiality and well-conserved nature of Nus factors among the prokaryotes make them potential targets for new drugs. The availability of the structures of each of the Nus factors will enable structure-based designs of drugs against these factors.

ACKNOWLEDGMENTS

We thank Dipak Dutta and Bibhusita Pani for reading the manuscript and helping in making the figures. We thank Irina Artsimovitch for providing us with the coordinates of the structural model of the EC-RfaH complex.

The research in the R.S. laboratory is supported by NIH grant TW06185, a Wellcome Trust Senior Research fellowship, and an extramural grant from the Department of Biotechnology, India. J.C. and G.M. are a University Grants Commission senior research fellow and a Department of Biotechnology junior research fellow, respectively.