What the Hel: recent advances in understanding rifampicin resistance in bacteria

Abstract Rifampicin is a clinically important antibiotic that binds to, and blocks the DNA/RNA channel of bacterial RNA polymerase (RNAP). Stalled, nonfunctional RNAPs can be removed from DNA by HelD proteins; this is important for maintenance of genome integrity. Recently, it was reported that HelD proteins from high G+C Actinobacteria, called HelR, are able to dissociate rifampicin-stalled RNAPs from DNA and provide rifampicin resistance. This is achieved by the ability of HelR proteins to dissociate rifampicin from RNAP. The HelR-mediated mechanism of rifampicin resistance is discussed here, and the roles of HelD/HelR in the transcriptional cycle are outlined. Moreover, the possibility that the structurally similar HelD proteins from low G+C Firmicutes may be also involved in rifampicin resistance is explored. Finally, the discovery of the involvement of HelR in rifampicin resistance provides a blueprint for analogous studies to reveal novel mechanisms of bacterial antibiotic resistance.


Introduction
Unlike in experimental settings in the laboratory involving single species, bacteria exist in Nature in communities.Within these communities, all types of interactions can be found, from symbiosis to indifference to open hostility.To survive and navigate this complex environment, bacteria must sense, interpret, and respond to numerous signals and protect themselves against various toxic compounds, such as antibiotics.
Many of the currently used antibiotics were discovered in Actinobacteria (Mast and Stegmann 2019 ), a phylum of mostly grampositive bacteria of great medicinal and economical importance.Rifampicin (synonym rifampin) was discovered in the mid-1960's, and was soon introduced into ther a peutic use (Sensi 1983 ).Rifampicin (and its deri vati ves) functions by binding to the β subunit (encoded by the rpoB gene) of bacterial RNA pol ymer ase (RN AP) in its DN A/RN A channel.When bound, rifampicin obstructs this channel and pr e v ents extension of RNA during transcription initiation beyond the first 2-3 nucleotides (nt), effectiv el y shutting down transcription, the first step of gene expression (Campbell et al. 2001, Lin et al. 2017 ).Conv ersel y, when RNA is longer than 3 nt, it blocks the rifampicin binding site.
Resistance to rifampicin arises due to mutations in amino acids (aa) involved in its binding to RNAP (Leehan and Nicholson 2021 ).These mutations adversely affect the activity of RNAP and this is counteracted by secondary mutations in RNAP that restore fitness (Kurepina et al. 2022 ).In many Nocardia species there are even two copies of rpoB , one copy ( rpoB2 ) dedicated to providing resistance to rifampicin (Ishikawa et al. 2006 ).Mor eov er, bacteria hav e de v eloped se v er al additional a ppr oac hes to deal with rifampicin, either by active efflux (Li et al. 2015, Machado et al. 2018 ) or by modifications such as phosphorylation, ADP-ribosylation, glycosylation, or hydroxylation (Tupin et al. 2010, Goldstein 2014, Stogios et al. 2016 ).These modifications alter the binding properties of rifampicin, pr e v enting its stable interaction with RNAP .Recently , a nov el mec hanism of bacterial r esistance to rifampicin was r eported, mediated by the HelD protein in Actinobacteria (Hurst-Hess et al. 2022, Surette et al. 2022 ).The authors proposed to refer to this protein as HelR in this phylum.In this r e vie w, the term 'HelD' is used for both HelD and HelR, regardless of the species, unless specified otherwise .T his r e vie w addr esses structur al and functional aspects of the mechanism of action of HelD/HelR proteins and discusses the so far poorly characterized mechanisms of their expr ession contr ol.

HelD-discovery & functional interaction with RNAP
HelD is a helicase-like ATP ase/GTP ase without a helicase function.It was first investigated with respect to homologous recombination in Bacillus subtilis ( Bsu ) (Carrasco et al. 2001 ).A decade later, it was detected as an RNAP associating protein in B. subtilis (Delumeau et al. 2011 ).Subsequent in vitro studies then established that HelD dissociates stalled/inactive complexes of RNAP fr om DNA, liber ating the template for further tr anscription and pr e v enting potentiall y deleterious tr anscription-r eplication collisions, thereby contributing to genome integrity maintenance (Wiedermanno va et al. 2014 ).T his function of HelD is synergistically enhanced by δ, a small subunit of RNAP specific for Firmicutes (Juang and Helmann 1994, Doherty et al. 2010, Motac k ov a et al. 2010, Rabatinova et al. 2013, Kuban et al. 2019 ), a phylum where B. subtilis belongs.HelD thus r epr esents one of se v er al mec hanisms that ensur e r emov al of RNAPs r oadbloc king DNA (Wiedermannov a and Kr asn y 2021 ).This functional r edundancy is pr obabl y reflected in the minor phenotypes of the absence of HelD-slo w er growth in the presence of elevated KCl concentration and prolonged lag phase after dilution of stationary phase cells into fresh medium (Wiedermannova et al. 2014 ).The latter phenotype is also consistent with another proposed function for HelD-storing inactive RNAPs in stationary phase (Pei et al. 2020 ).The sequestered RNAPs are reactivated when the cells encounter more advantageous conditions, reminiscent of similar roles of 6S and Ms1 sR-NAs that sequester the primary σ factor-containing RNAP holoenzyme or the RNAP core (subunit composition α 2 ββ' ω), r espectiv el y (W assarman 2018 , V ank ov a Hausner ov a et al. 2022 ).

HelD-structural aspects
In 2020, the structure of HelD was determined in complex with RN AP b y cry o-EM.Thr ee structur al studies of HelD and RNAP were published side-by-side at that time, two from B. subtilis and one from Mycobacterium smegmatis ( Msm ) (Kouba et al. 2020, Newing et al. 2020, Pei et al. 2020 ).These studies r e v ealed a unique type of interaction betw een RN AP and a pr otein factor.HelD fr om both species penetrates both the primary channel (where DNA binds) and secondary channel (where NTPs enter the active site), embracing RNAP.This causes large conformational changes in RNAP.T he cla ws of the two largest subunits , β and β', that form the primary channel become wide open, which, in turn, together with the obstructing presence of HelD on RNAP and widening of the RNA exit channel, is incompatible with the presence of nucleic acids (Fig. 1 ).
HelD pr oteins fr om both or ganisms, Bsu and Msm , contain two ATPase RecA-like domains and the Walker motif, which indicates ATP/GTP binding (Walker et al. 1982 ).Indeed, HelD proteins bind and hydr ol yze ATP and/or GTP, whic h leads to conformational changes in HelD (K o val et al. 2019 ).Nevertheless, it is not clear at which step the hydrolysis occurs and the details of these conformational changes are not defined yet.In all available RNAP-HelD complexes, HelD is present without a nucleoside triphosphate (NTP).Addition of ATP or its nonhydr ol yzable analogue then induces dissociation of HelD from RNAP in B. subtilis (Pei et al. 2020 ), arguing that the changes required for HelD release are caused by binding but not hydr ol ysis of ATP.HelDs from both organisms also possess the clamp opening (CO) domain that interacts with β', holding the DNA clamp wide open, and the N-terminal domain that inserts into the secondary channel of RNAP (Fig. 2 ).
HelDs from the tw o organisms, ho w ever, differ in some aspects of their structures (Fig. 3 ), and, consequently in their interaction with RNAP.In the case of Bsu HelD, the N-terminal domain r eac hes deep into the secondary channel, interfering with the active site of RNAP.In the case of Msm HelD, the N-terminal domain does not r eac h as deep but still immobilizes the trigger loop (TL) as in Bsu HelD (Newing et al. 2020, Pei et al. 2020 ).The TL is a flexible component of RNAP that cycles between 'open' and 'closed' states with respect to nucleoside triphosphate (NTP) entry to the active site (Mazumder et al. 2020 ).The interaction of the N-terminal domain of HelD with the TL freezes the TL in an open-like state, whic h interfer es with the NTP addition cycle (Kouba et al. 2020 ).Bsu HelD then lacks the primary channel (PCh) loop by which Msm HelD r eac hes fr om inside the primary channel directly to the active site.
The structural differences are also reflected in terminology: Bsu -like HelDs are termed Class I and Msm -like HelDs Class II (Larsen et al. 2021 )   wher e some gener a, suc h as Streptom yces, Nonomuraea , and Frankia , e v en contain se v er al Class II HelD par alogs (Larsen et al. 2021 ).Figur e 3 illustr ates the pr ominent structur al differ ences between HelD pr oteins fr om both classes.Very r ecentl y, Class III HelD pr oteins were identified bioinformatically by Larsen et al. in Gramnegativ e Delta pr oteobacteria (Larsen et al. 2021 ).A hallmark of Class III HelDs is an extended motif within the N-terminal domain, DWR X [A/S]P (extended by one aa, X ), and the absence of the PCh loop.Furthermore, their N-terminal domain, expected to bind into the secondary channel of RNAP, is shorter than this domain in Class I HelDs and longer than in Class II HelDs.Examples of bacterial species with identified helD genes (Class I, Class II, and Class III) are in Table 1 .

HelD-involvement in rifampicin resistance-discovery
Already in 2004, Hutter et al .reported that the Bsu yvgS (synonym for helD ; it was a gene of unknown function then) promoter region was inducible by rifampicin (Hutter et al. 2004 ).A followup study confirmed this result and sho w ed the inducibility of the yvgS pr omoter r egion also by another RNAP-targeting compound, Table 1.Classes of HelD proteins with re presentati ve bacterial species.Class I are low G + C Gram-positive bacteria and Gram-negative Bacteroidia , Class II are high G + C Gr am-positiv e bacteria.Class III are found in Gram-negative bacteria.For a more detailed list and phylogeny see (Larsen et al. 2021 ).Mycobacterium abscessus is a distant r elativ e of Mycobacterium tuberculosis and Mycobacterium leprae but in comparison to these pathogens it is a fast-growing bacterium (Maurer et al. 2014 ).M. abscessus can cause infections in humans .T hese infections are usually limited to the skin and the soft tissues under the skin, although in persons with v arious c hr onic lung diseases, such as cystic fibrosis it can also cause serious lung infections.Notably, M. abscessus is known for its resistance to rifampicin, unlike M. tuberculosis and M. leprae that do not usually contain the helD gene (see Table 1 ).This absence may be due to the environment that these human pathogens inhabit, where they have not been exposed to these antibiotics as m uc h ov er a long evolutionary time frame.Conv ersel y, S. venezuelae is a soil-dwelling or ganism wher e it faces the c hallenge fr om antibiotic pr oducers (Chater 2016 ).Suc h organisms typically possess multiple lines of defense against antibiotics to enhance their chances of survival in the hostile environment.

Mechanisms of HelD-mediated rifampicin resistance
Both recent studies (Hurst-Hess et al. 2022, Surette et al. 2022 ) sho w ed that deletion of the HelD-encoding gene results in increased sensitivity to rifampicin in the tested actinobacterial species .T his correlates with the ability of HelD to dissociate stalled RNAPs (Kouba et al. 2020 ).Stalled RNAPs are also those bound to the pr omoter, arr ested by rifampicin.When the rifampicin concentration is not saturating and HelD dissociates rifampicin-arr ested RNAPs fr om pr omoters, rifampicin-fr ee RNAP molecules can initiate tr anscription fr om these promoters.Indeed, Hurst-Hess et al .sho w ed that HelD can do exactly that (Hurst-Hess et al. 2022 ).Although this may already seem to provide significant protection against rifampicin, Class II HelDs can do e v en better.
Using a rifampicin photoaffinity probe (RPP) linked to biotin, Surette et al. sho w ed that S. venezuelae HelD decreases binding of RPP to RNAP , and, importantly , it can also dissociate rifampicin fr om RNAP (Sur ette et al. 2022 ).This dissociation seems to depend on the HelD PCh loop, and mutations of two amino acid residues at the tip of this loop to alanines (Glu496 and Asp497, numbering according to M. abscessus ; Glu484 and Asp485 in Msm HelD, MSMEG_2174 gene) compromise the ability of HelD to render the cell rifampicin-resistant even though this HelD mutant is still capable of binding to RNAP, and dissociating stalled RNAPs Binding of the Bsu HelD N-terminal domain (shown as sticks and coil with carbon in magenta) in the RNAP secondary channel and active site (indicated with Mg 2 + ) deforms the expected rifampicin binding poc ket, whic h is indicated with the arrow; aa residues closest to the rifampicin pocket are shown as sticks.Rifampicin is shown as in panel A with increased transparency for clarity.The Bsu RNAP chains are shown in secondary structure r epr esentation ( Bsu elongation complex colored pale gray, HelD complex colored teal) with selected aa residues shown as sticks .P otential atomic clashes between rifampicin and its deformed binding site are indicated with the red square brackets .T he graphics was created using PyMOL.
from DNA (Hurst-Hess et al. 2022 ).This also suggests that dissociation of rifampicin from RN AP b y HelD and not dissociation of the RNAP-rifampicin complex from DNA is the main mode of how HelD provides resistance to rifampicin.
A comparison of the Msm RNAP-HelD complex with the Mycobacterium tuberculosis ( Mtu ) RNAP-rifampicin complex r e v eals that the Msm HelD PCh loop affects the geometry of the rifampicin binding pocket (Fig. 4 A).The PCh loop tip is within 3.5 Å of the RNAP rifampicin binding site and Asp485 can dir ectl y inter act with the rifampicin binding poc ket.Mor eov er, aa r esidues of the PCh loop that follow Asp485 in the peptide chain deform the pocket so that the presence of both-rifampicin and HelD-at the same time would likely lead to serious atomic clashes.Furthermor e, Ar g456 of the Msm β subunit (Arg461 in M. abscessus ), corresponding to Arg465 in the Mtu RNAP-rifampicin complex (Lin et al. 2017 ), which forms a direct hydrogen bond to the antibiotic, is div erted fr om the rifampicin binding poc ket in the Msm RNAP-HelD complex and forms a hydrogen bond to Gln490 of Msm HelD.This altered geometry of the rifampicin pocket likely leads to a decreased affinity for rifampicin binding.Additionally, ATP binding/hydr ol ysis to/by HelD may elicit additional conformational changes in the rifampicin poc ket, pr omoting dissociation of the antibiotic from RNAP.Based on the relatively small extent of the conformational changes induced in the rifampicin binding pocket by HelD binding, it is unlikely that these c hanges ar e sufficientl y long-lived to provide additional protection after HelD dissociation.Such a phenomenon was proposed for tetracycline ribosomal protection proteins (Wilson et al. 2020 ).
Mutants in the Walker motif of M. abscessus HelD were also defective in eliciting rifampicin resistance but they seemed to have a significantly decreased ability to bind RNAP, probably due to a compromised conformation (Hurst-Hess et al. 2022 ).Taken together, this classifies Class II HelD proteins as Type II tar get pr otection, an incr easingl y mor e a ppr eciated mec hanism of antibiotic r esistance, wher e the tar get pr otection pr oteins induce conformational changes in the target that allosterically dissociate the antibiotic from the target (for review see (Wilson et al. 2020 )).
The pr e vious par a gr a phs of this section describe studies in high G + C Actinobacteria (Class II HelDs or HelRs).Surette et al .also tested the effect of the absence of a Class I HelD from B. subtilis on rifampicin MIC and detected no difference compared to wt (Surette et al. 2022 ).They concluded that Class I HelD proteins are likely not involved in rifampicin resistance.For Class III HelDs, no experimental evidence exists about their potential involvement in rifampicin resistance.

Mechanisms of rifampicin-induced expression of HelD
Induction of transcription in bacteria by sub-MIC rifampicin has been r eported pr e viousl y (Yim et al. 2013 ), perha ps most notabl y for rpoBC genes (Ho w e et al. 1982, Zhu et al. 2018 ), encoding the two largest subunits of RNAP, β and β'.This induction was associated with increased resistance to rifampicin.Ho w ever, upregulation of rpoBC mRNA in M. smegmatis (2x ↑ ) was not as pr onounced as upr egulation of helD mRNA in M. abscessus (25x ↑ )  (Zhu et al. 2018, Hurst-Hess et al. 2022 ).Ne v ertheless, it was shown that the Msm rpoBC operon is transcribed from two promoters, the first (more upstream relative to the genes) promoter is less active and interferes with transcription from the second promoter.When sub-MIC rifampicin is present, it preferentially inhibits tr anscription fr om the first pr omoter and this incr eases tr anscription fr om the second promoter (Zhu et al. 2018 ).This implies that the two pr omoters ar e differ entiall y susceptible to binding of RNAP with/without rifampicin, a phenomenon that will be of interest to define in detail.
Regulation of Class II HelD expression may also depend on two promoters as in M. smegmatis (Martini et al. 2019 ) promoter 1 and 2 were identified to drive expression of the helD gene (Fig. 5 A).The experimental evidence is available only in the absence of rifampicin when expression of HelD is relatively low, which is sufficient for its role in maintenance of genome integrity.Based on homology, the 5' UTR of M. abscessus helD displays the same promoter arc hitectur e .T he stimulation of expression of Class II HelD by rifampicin is controlled at the tr anscriptional r ather than tr anslational le v el (Hurst-Hess et al. 2019, Hurst-Hess et al. 2022 ).The RAE, r equir ed for the stimulation, is ∼20 bp downstream from the transcription start of the σ A / σ B -dependent promoter 1 and overlaps with the −35 region of the σ A / σ B -dependent promoter 2. The RAE consists of a 19 bp palindromic sequence (Spanogiannopoulos et al. 2014 ).To such cis -regulatory palindromes, transcription factors are known to bind (Narlikar and Hartemink 2006 ), either as homodimers or heterodimers, in the latter case offering combinatorial gene expression regulation.
When the upstream half of the RAE was deleted, abolishing the palindrome, the induction of helD by rifampicin was lost (Hurst-Hess et al. 2022 ).This loss of inducibility could be due to two reasons.First, the partial deletion of the RAE also r emov ed the -35 region of promoter 2, likely compromising its activity.Second, association of an unknown, rifampicin-binding activator with the RAE was pr e v ented.
In B. subtilis , the helD promoter activity is clearly inducible by rifampicin e v en though it does not contain the RAE (Hutter et al. 2004, Urban et al. 2007 ).On the other hand, the involvement of HelD in rifampicin resistance has not been detected and Class I HelDs lack the PCh loop (Surette et al. 2022 ).It is difficult to reconcile these results and we speculate that perhaps using more sensitive assays ( e.g.time kill), it may turn out that Bsu HelD as well as other Class I and possibly also Class III HelDs are involved in rifampicin resistance albeit not as str ongl y as in Actinobacteria .
The k e y function of the PCh loop in rifampicin resistance may in Class I HelDs instead be mediated by amino acid residues in the N-terminal domain that dir ectl y inter act with amino acid residues of the rifampicin binding pocket (Pei et al. 2020 ).Fig. 4 B illustrates this situation, showing the potential of Class I HelD to interfere with rifampicin binding.HelD forces part of β, most notably E521, into the rifampicin binding pocket.
Potentially, the δ subunit of RNAP may also be involved as its flexible C-terminal domain may extend into the DN A/RN A channel.Regardless of whether Bsu HelD has some role in rifampicin resistance or not, the experimentally demonstrated rifampicindependent induction of its expression is mediated by a yet to be determined mechanism.

Model of the HelD cycle
The current knowledge maps many, but not all, of the steps of the HelD cycle in transcription and rifampicin resistance.Also, ther e ar e differ ences with r espect to what types of RNAP complexes HelD binds between B. subtilis (Class I) and M. smegmatis (Class II).Bsu HelD was shown to bind the RNAP core (Wiedermannova et al. 2014 ) while Msm HelD was found to bind both the RNAP core and holoenzyme containing σ A and RbpA (Kouba et al. 2020 ).RbpA is an Actinobacteria -specific transcription factor that increases the rate of open complex formation between RNAP σ A and promoter DNA (Jensen et al. 2019 ).Inter estingl y, RbpA was pr e viousl y r eported to affect rifampicin resistance in Actinobacteria (Newell et al. 2006 ).This effect seems to stem from its ability to boost transcription in general and not from affecting rifampicin interaction with RNAP (Hu et al. 2012 ) although a study by Verma et al. reported that in an in vitro heterologous system Mtu RbpA was capable of rescuing rifampicin-induced transcription inhibition of Msm RNAP (Verma and Chatterji 2014 ).
Both Class I and II HelDs can bind to stalled elongation complexes and dissociate them (Kouba et al. 2020, Hurst-Hess et al. 2022 ).The binding likely occurs in a nucleotide free state and it is belie v ed to be initiated by the N-terminal domain of HelD.Binding of ATP/GTP is r equir ed for dissociation of Class I HelD from RNAP (Pei et al. 2020 ).This effect of ATP/GTP binding on Class II HelD r elease fr om RNAP has not been tested yet.This dissociation may be further stimulated by interaction with DNA but experimental e vidence is lac king.We speculate that in the cytoplasm, HelD then hydr ol yzes ATP, r esetting the molecule for another round of interactions.With respect to rifampicin, Class II HelD w as sho wn to dissociate rifampicin-stalled RNAPs from DNA and also rifampicin from RNAP in the absence of nucleic acids.A simplified model of the Class II HelD cycle is depicted in Fig. 5 B-E.We note that this model a ggr ees in most aspects with the model presented in (Hurst-Hess et al. 2022 ).The main difference is the mode of HelD r elease fr om RNAP with r espect to ATP hydr ol ysis.
Finall y, when ATP/GTP le v els ar e r elativ el y low in the cell, suc h as during nutritional starvation in stationary phase, the absence of ATP/GTP binding likely contributes to sequestration of RNAP by Class I HelD, k ee ping it in an inacti ve form until conditions impr ov e (Pei et al. 2020 ).In the case of Class II HelDs, this scenario has not been explored.

Concluding remarks
Bacterial HelD proteins (all Classes) interact in a unique manner with RNAP.Some aspects of this interaction are yet to be specified, such as the sequence of e v ents leading to, and during HelD/HelR r elease fr om RNAP.Class II HelD (HelR) proteins then r epr esent a ne w mode of rifampicin r esistance.Inter estingl y, r esistance to another RNAP-binding antibiotic fidaxomicin is not affected by Class II HelD proteins (Surette et al. 2022 ) even though HelD binding to RNAP dilates the RNA exit channel (Kouba et al. 2020 )-the fidaxomicin binding site (Bo y aci et al. 2018 ).Future studies will be r equir ed to elucidate the exact mechanism by which rifampicin is released by Class II HelDs from RNAP.It also remains to be explored in more detail whether Class I and III HelDs possibly also play some role in rifampicin resistance.Mor eov er, r esistance to rifampicin might also be influenced by the presence of other interacting microbes (Bottery et al. 2021 ); deciphering such effects within complex communities may yield medicinall y r ele v ant insights.Last but not least, addressing the mechanisms of rifampicin-dependent expression control then is of prime interest and may reveal novel types of transcriptional regulation.
Finall y, the discov ery of rifampicin r esistance by Class II HelD pr oteins r e v eals how r elativ el y little w e kno w about e v en wellstudied antibiotics e v en though the effects of sub-MIC of antibiotics have been long recognized (Davies et al. 2006 ).'Omic' appr oac hes similar to those used pr e viousl y with rifampicin (Giddey et al. 2017, Hurst-Hess et al. 2019, Surette et al. 2022 ) may be applied to other antibiotics when their sub-MIC may lead to upr egulation of pr e viousl y unidentified pr oteins involv ed in antibiotic resistance, and to a better understanding of risks posed by hormetic dose responses of bacteria to environmental contamination by antibiotics due to their overuse (Iavicoli et al. 2021 ).

F igure 1 .
Msm RN AP and its interaction with nucleic acids (A) or HelD (B).(A) Initiation complex of Msm RN AP (PDB id 5VI5, Hubin et al. 2017 ) in secondary structure representation, color coding: RNAP α1, pale y ello w; α2, wheat; β, teal; β', blue; ω, pink; σ , dark purple; RbpA, bright y ello w; template DNA strand, bright orange; non-template DNA strand, bright green; nascent RNA, red; the active site of RNAP is marked with modeled Mg 2 + -hotpink sphere.(B) Complex of the Msm RNAP core with HelD (PDB id 6YY S, K ouba et al. 2020 ) in secondary structure representation, color coding as in (A) with the addition of HelD in or ange; Mg 2 + experimentall y localized in the active site-hotpink sphere .T he graphics were created using PyMOL (The PyMOL Molecular Gr a phics System, Version 2.0, Schrödinger, LLC).
. Class I HelDs are found in the low G + C Grampositive bacteria and gram-negative Bacteroidia , in the latter case likel y acquir ed by horizontal gene tr ansfer fr om anaer obic gut Clostridiales due to environmental coexistence of these organisms.Class II HelDs are present in the high G + C Gram-positive bacteria

Figure 3 .
Figure 3.Comparison of Msm and Bsu HelD structures.The molecules are superimposed as in Fig. 2 , i.e. in equivalent positions when binding RNAP.Both molecules are in secondary structure representation.Msm HelD is colored orange, Bsu HelD purple.Mg 2 + present in the active site of Msm RNAP is shown as magenta sphere.Amino acids of the active site coordinating it are shown as sticks with carbon in blue .T he most profound differences between HelD from Msm and Bsu are marked.The graphics was created using PyMOL.

Figure 4 .
Figure 4. HelD causes conformational changes in the RNAP rifampicin binding site.(A) Structural comparison of the rifampicin binding site of the Mtu RNAP-rifampicin complex (PDB id 5UHC, Lin et al. 2017 ) with the same site from the Msm RNAP-HelD complex (PDB id 6YY S, K ouba et al. 2020 ) is shown.Binding of the Msm HelD PCh loop (shown as sticks with carbon in red and in red secondary structure representation) in the RNAP primary channel and active site deforms the rifampicin (RIF) binding pocket.Rifampicin is shown with carbon colored pale orange.Rifampicin binding pockets from both structures are shown in secondary structure representation ( Mtu colored pale gray, Msm colored teal) with the selected residues shown as sticks .T he indicated hydrogen bonds and Mg 2 + coordination distances are given in Å. Possible atomic clashes between rifampicin and its deformed binding site and HelD are hinted with the red 'wave' symbols.(B) Structural comparison of expected rifampicin binding in Bsu RNAP (PDB id 6WVJ, Newing et al. 2020 ) and the same site in the complex of Bsu RNAP with HelD (PDB id 6ZFB (Pei et al. 2020 )) superimposed by the main chain of β.Binding of the Bsu HelD N-terminal domain (shown as sticks and coil with carbon in magenta) in the RNAP secondary channel and active site (indicated with Mg 2 + ) deforms the expected rifampicin binding poc ket, whic h is indicated with the arrow; aa residues closest to the rifampicin pocket are shown as sticks.Rifampicin is shown as in panel A with increased transparency for clarity.The Bsu RNAP chains are shown in secondary structure r epr esentation ( Bsu elongation complex colored pale gray, HelD complex colored teal) with selected aa residues shown as sticks .P otential atomic clashes between rifampicin and its deformed binding site are indicated with the red square brackets .T he graphics was created using PyMOL.

Figure 5 .
Figure 5. Class II HelD promoter region architecture and a model of the Class II HelD cycle.(A) Alignment of promoter sequences for helD gene in M. smegmatis (MSMEG_2174) and M. abscessus (MAB_3189c).The transcription start site (TSS) for P1 and P2 of M. smegmatis were experimentally determined (Martini et al. 2019 ); for M. abscessus the TSS are putative, based on homology and distance from the RAE.The TSS and putative −35 and −10 consensus elements ( σ A / σ B , (Zhu et al. 2017 )) are highlighted with grey background.Conserved Gs (immediately upstream of −10, underlined) are typical for mycobacterial promoters (Zhu et al. 2017).The RAE is in the rectangle, the palindrome indicated with arrows.(B) Rifampicin can bind RNAP, RNAP σ A .Rifampicin blocks transcription at an early stage.Elongating RNAP can be stalled on DNA by the presence of an obstacle.(C) HelD can dissociate RNAP from nucleic acids and rifampicin.The light green part of HelD indicates that this part of HelD is obscured by RNAP.(D) Binding of A TP/GTP (A TP is shown onl y for simplicity) induces dissociation of HelD fr om RNAP.Liber ated RNAP can r e-enter the tr anscription cycle.(E) HelD hydr ol yzes ATP/GTP and can interact with RNAP again.HelD is also generated by translation of helD mRNA, especially after induction of helD transcription by rifampicin.For more details, see text.