MucD regulates alginate biosynthesis through the proteolytic control of AlgX and AlgK in Pseudomonas aeruginosa Yujun Jiang, Xin-Fu Yan, Rya Ero, Chao Wang, Kanaga Sabapathy, Yong-Gui Gao, Ann M. Stock Abstract Pseudomonas aeruginosa is an opportunistic human pathogen capable of infecting a wide range of tissues and organs. Its persistence during chronic infection is strongly associated with biofilm formation, which depends on extracellular polysaccharides such as alginate. The HtrA-like periplasmic serine protease MucD is a key regulator of bacterial virulence, stress response, and alginate production, yet its molecular mechanism has remained largely unclear. Here, we discovered the alginate acetylation and export proteins AlgX and AlgK as MucD substrates, and characterized their degradation by mass spectrometry and bioinformatic analysis. We further determined the cryo-EM structure of MucD bound to an AlgK-derived substrate peptide, offering atomic insights into MucD oligomerization assembly, substrate recognition, and specificity. Together with structure-guided mutagenesis and biochemical assays, our results revealed that MucD proteolytic activity is governed by an equilibrium between a resting 12-mer and an active trimer. Crucially, we demonstrate that MucD represses alginate biosynthesis post-translationally, in addition to its previously implicated role in transcriptional regulation. These findings define a distinct activation mechanism and regulatory function for MucD and provide new insight into bacterial HtrA-like serine proteases. The persistence of Pseudomonas aeruginosa during chronic infection relies on alginate production, a process regulated by the serine protease MucD. This study identifies the alginate acetylation and export proteins AlgX and AlgK as MucD substrates and shows that MucD proteolytic activity is regulated by oligomerization. Introduction Pseudomonas aeruginosa is a major opportunistic human pathogen and a frequent cause of nosocomial infections, particularly in immunocompromised patients []. Its pathogenicity is supported by the production of diverse virulence factors, including extracellular polysaccharides (EPS). EPS are key components of bacterial biofilm, which shields bacteria from antibiotics, oxidative stress, and host immune defenses during chronic infections []. Alginate, together with the EPS components Pel and Psl, plays an important role in biofilm formation and maturation. Alginate is synthesized through a synthase-dependent pathway encoded by 12-gene alg operon []. Transcription of the alg operon is tightly controlled by the sigma factor AlgU/T, and mutations affecting AlgU/T or its regulatory proteins are frequently found in clinical alginate-overproducing (mucoid) isolates []. Protease-mediated protein quality control is essential for cellular homeostasis, particularly under environmental stress []. The high-temperature requirement A (HtrA) family of serine proteases is conserved from bacteria to humans and contribute to stress adaptation by degrading misfolded or damaged proteins []. HtrA-like proteases typically contain a structurally conserved protease domain and one or two PDZ (post synaptic density protein 95 (PSD95), Drosophila disc large tumor suppressor (Dlg1), and zonula occludens-1 protein (zo-1)) domains [], which mediate protein-protein interactions and substrate binding []. In gram-negative bacteria, HtrA-like proteases have attracted interest as potential antimicrobial targets because of their important physiological roles, conserved architecture, and defined substrate-binding sites []. For example, small-molecule inhibition of HtrA-like protease DegS suppresses stress-response activation and antibiotic resistance in E. coli []. DegP, another HtrA-like protease in gram-negative bacteria, is central to periplasmic protein homeostasis []. Notably, its activity is tightly controlled by oligomerization: substrate binding converts an inactive hexameric resting state into larger, proteolytically active 12-mer and 24-mer assemblies []. The mucD gene is located within an autoregulated operon comprising algU, mucA, mucB, mucC, and mucD in P. aeruginosa []. MucD is an HtrA-like periplasmic protease and DegP homologue found in the genus Pseudomonas []. It contains a protease domain followed by two C-terminal PDZ domains (PDZ1 and PDZ2), and has been implicated in bacterial response to heat stress and reactive oxygen species [,]. Loss of mucD attenuates virulence in infection models of Arabidopsis, C. elegans, and mice [,]. MucD also negatively regulates alginate biosynthesis at the transcriptional level by indirectly repressing AlgU/T activity []. Affinity pull-down studies further suggested that MucD could form a periplasmic complex with the terminal acetyltransferase AlgX and tetratricopeptide repeat (TPR)-rich scaffold protein AlgK []. Both AlgK and AlgX are required for alginate chain maturation, modification, and secretion []. Despite the established roles of MucD in bacterial stress response, virulence, and alginate biosynthesis, its proteolytic function, physiological substrates, and structural basis of substrate recognition remain poorly understood. To date, the only reported MucD structure is that of a truncated P. syringe MucD lacking the PDZ2 domain []. Here we identified AlgK and AlgX as MucD substrates, determined the cryo-EM structure of substrate-bound full-length MucD, and characterized its proteolytic properties, oligomerization mode, substrate recognition and specificity, as well as physiological role in alginate biosynthesis. Together, our results define the functional and structural characteristics of MucD and offer atomic insights into HtrA-like protease-mediated proteolytic regulation in P. aeruginosa. Results Alginate biosynthesis pathway proteins AlgK and AlgX are substrates of MucD protease Among alginate biosynthesis proteins, AlgK acts as a structural scaffold, tethering the inner-membrane alginate polymerase to the outer-membrane pore []; AlgX is an essential periplasmic acetyltransferase that O-acetylates the polymer, a modification required for alginate gel-like properties and resistance to host defenses []. Together, they form the AlgKX complex, which ensures proper alginate maturation and its transit through the periplasm prior to secretion []. A previous study reported a large periplasmic complex involving MucD, AlgK, and AlgX [], but whether AlgK and AlgX are MucD substrates remained unknown. To address this question, MucD was expressed and purified from the wild-type P. aeruginosa PAO1 strain. The initial size-exclusion chromatography (SEC) profile showed that MucD primarily exists in two oligomeric states (). Isolation and re-application of two oligomeric populations to analytical SEC separately further identified them as a trimer and a 12-mer (main population), based on molecular weight (MW) calibration ( and ). Analytical SEC also indicated that the MucD 12-mer and trimer exist in a dynamic equilibrium in solution. Furthermore, no convincing additional protein co-eluting with MucD was detected based on SDS-PAGE () or MALDI-TOF/MS (). We next tested whether AlgK and AlgX are MucD substrates. Notably, degradation assays demonstrated that both the MucD 12-mer and trimer are able to completely degrade full-length AlgK and AlgX within 2 h, with no significant difference in the degradation efficiency between the two oligomeric states ( and –). Additionally, no partially degraded fragments were detected by SDS–PAGE, suggesting complete substrate turnover. Substrate binding converts DegP from a resting hexamer into active higher-order oligomers, including 12-mer and 24-mer assemblies []. We therefore asked whether MucD undergoes a similar substrate-dependent change. AlgK was incubated with either the MucD trimer or 12-mer for 15 min or 2 h; under these conditions, AlgK was partially degraded at 15 min and fully degraded after 2 h, respectively ( and ). Analytical SEC results showed that a brief 15-min incubation of the MucD 12-mer with AlgK caused limited substrate degradation, accompanied by a modest decrease in the 12-mer population and an increase in the trimer population. This shift did not revert after prolonged incubation, even when degradation was complete (). A similar oligomerization behavior was observed when AlgK was incubated with the MucD trimer (). Taken together, these results suggest that substrate binding shifts the MucD oligomeric equilibrium toward the trimer, which appears to be the preferred state during proteolysis. AlgK and AlgX form an AlgKX complex, which is essential for alginate biosynthesis in vivo, and the structure of this complex was recently determined in P. putida []. To test whether the AlgKX complex is also a proteolytic target of MucD, we first reconstituted the complex by co-incubating purified AlgK and AlgX proteins from P. aeruginosa, followed by analytical SEC analysis (). The dominant peak corresponding to the AlgKX complex had an apparent MW of approximately 110 kDa, in line with the theoretical MW of 105.1 kDa for a 1:1 AlgK-AlgX heterodimer. Subsequently, our degradation assays revealed that the AlgKX complex was resistant to MucD proteolysis, with little degradation after 8 h and substantial degradation only after 24 h of incubation (). In contrast, isolated AlgK and AlgX were completely degraded by MucD within 2 h of incubation ( and –). These data indicate that AlgK and AlgX are MucD substrates as individual proteins, whereas the assembled AlgKX complex is protected from MucD-dependent degradation. High-resolution analysis of MucD-dependent proteolytic products of AlgK and AlgX To better understand how MucD recognizes and processes its substrate, we combined Liquid Chromatography-Mass Spectrometry (LC-MS) with bioinformatics analysis to identify MucD-dependent degradation products of AlgK and AlgX. Briefly, AlgK or AlgX was incubated with MucD for 8 h to promote extensive proteolysis. The resultant degradation peptides were then separated from intact proteins using a 10-kDa molecular weight cutoff (MWCO) concentrator prior to LC-MS analysis. MS data were further processed using the “Utilities for Mass Spectrometry Analysis of Proteins” (UMSAP) software []. In total, we detected 381 peptides matching AlgK, corresponding to 90.1% sequence coverage, and 321 peptides matching AlgX, corresponding to 94.7% sequence coverage (). The average peptide lengths were 13 and 14 amino acids for AlgK and AlgX, respectively. These results are consistent with previous degradation assays ( and –) and confirm extensive MucD-mediated proteolysis of both substrates. The background peptides generation was negligible, as only 39 peptides matching AlgK (15.6% sequence coverage) and 38 peptides matching AlgX (14.3% sequence coverage) from substrate-only incubations were detected by LC-MS (). Proteolysis cleaves a peptide bond, generating two shorter fragments. Under standard protease nomenclature, the substrate residue at the N-terminal side of the scissile bond is designated as the P1 residue, while the residue at the C-terminal side is designated as the P1’ residue []. UMSAP also calculates the relative frequency of cuts (RFC) at each P1 residue, providing a quantitative measure of cleavage preference for a protease. For AlgK, 212 cleavage sites were mapped to the primary sequence, and 17 sites with RFC value greater than 9 were classified as high-RFC (hRFC) residues (). Given that no experimental structure is currently available for P. aeruginosa AlgK, we mapped these hRFC residues onto an AlphaFold-predicted AlgK model. Only about half of the hRFC residues were surface-exposed (). Two clusters of hRFC residues (A100, R101, and G103; L315, G317, and L319) were located in close spatial proximity, which may reflect limits in assigning the exact N- or C-terminal ends of polypeptide fragments. AlgX showed a similar pattern when hRFC residues were mapped onto the AlgX structure (PDB: 4KNC) []. Among 210 detected cleavage sites, eight were classified as hRFC residues (), and only five were surface-exposed (). Recent work on the HtrA protease HTRA1 showed that degradation of folded substrates can proceed sequentially: initial cleavage at surface-exposed sites induces local structural relaxation, followed by unfolding of early degradation products and exposure of buried cleavage sites []. MucD may use a similar mechanism to fully degrade native AlgK and AlgX, despite the burial of many cleavage sites. This model also helps explain why the AlgKX complex resists MucD proteolysis ( and ). Particularly, AlgK Q366 and AlgX V212 are hRFC residues located at the complex interface ( and ). These residues are surface-exposed in the isolated proteins and may participate in initial MucD cleavage events, but become buried upon AlgKX complex formation and are therefore shielded from proteolysis (, and ). Previous studies suggested that DegP preferentially cleaves after hydrophobic P1 residues with short side chains and shows limited specificity across P2-P5 positions, which often contribute to substrate recognition and proteases binding [,]. We therefore analyzed MucD sequence specificity by calculating amino acid distributions from P5 to P5′ using UMSAP. Overall, MucD showed limited sequence specificity for either AlgK or AlgX (), indicating that MucD acts as a protease that recognizes substrate conformation in addition to primary sequence. In particular, short side-chain nonpolar residues, such as alanine, valine, and leucine, were enriched at P1 position and accounted for more than 50% of P1 residues in both AlgK and AlgX degradation products (). Given that nonpolar residues are frequently buried in folded proteins, this enrichment is consistent with our observation that approximately half of the hRFC residues in AlgK and AlgX are not surface exposed. It also supports a model in which initial cleavage exposes buried nonpolar sites, enabling complete substrate degradation by MucD. UMSAP also applies a χ2 test to determine whether the observed P5-P5’ residue distributions differ significantly from a theoretical null distribution representing a protease with no positional selectivity. This null distribution is primarily based on the overall amino acid abundance of the substrate []. For both AlgK and AlgX, the analysis revealed statistically significant selectivity across P5-P1 positions (). For example, alanine, leucine, and valine were the three most enriched residues at the P1 site for both substrates. In contrast, P3′ and P4′ showed significant selectivity for AlgK but not for AlgX. Collectively, the χ2 test confirms the presence of specific, nonrandom amino acid patterns at the P5–P1 sites for both substrates. These distinct sequence profiles likely contribute to defining the substrate characteristics recognized by MucD. Structure of MucD in complex with its substrate polypeptide from AlgK To better understand the oligomeric state of MucD and the structural basis of substrate recognition, we initially sought to determine the structure of wild-type MucD, but did not succeed despite extensive efforts. We therefore adopted a strategy commonly used for structural studies of proteases, including the HtrA-like protease DegP [] and the AAA protease FtsH [], by generating a catalytically inactive variant. Specifically, we purified MucD(S217A), in which the catalytic serine is replaced by alanine, as the structural target. The SEC profile of MucD(S217A) resembled that of wild-type MucD, with both 12-mer and trimer populations present (). Protein degradation assays confirmed that MucD(S217A) is protease-deficient, as neither AlgK nor AlgX was degraded after 4 h of incubation (), in contrast to the wild-type MucD ( and –). Analytical SEC analysis further demonstrated that these two oligomeric states remain in dynamic equilibrium in solution ( and ). Interestingly, incubation of MucD(S217A) 12-mer with AlgK reduced the trimer population and promoted formation of 24-mer oligomers, which is distinct from that of the wild‐type MucD ( and ). SDS‐PAGE analysis of the 24-mer and 12-mer peaks showed co-elution of MucD with AlgK, suggesting that 24-mer formation is likely driven by substrate interaction (). Next, we asked whether the interaction between MucD(S217A) and its substrates can be resolved by cryo-EM. To improve sample homogeneity for structural study, we examined AlgK degradation products and selected several polypeptides containing hRFC P1 residues as candidate substrate peptides (). These peptides were chemically synthesized and tested their effect on the MucD(S217A) SEC profile. Eventually, we found that co-incubation of MucD(S217A) with AlgK369-388 (AlgK residues 369-388, N’-VDHLILAARAGQASADMALA) similarly induced 24-mer formation (), indicating that AlgK369-388 binds to MucD. Cryo-EM analysis then yielded structures of the MucD(S217A) 12-mer and 24-mer at resolutions of 4.6 Å and 3.5 Å, respectively ( and , ). The 12-mer and 24-mer complexes have overall diameters of ~155 Å and 190 Å, as well as enclosing an internal cavity of approximately 75 Å and 110 Å in diameter, respectively, dimensions comparable to the corresponding DegP oligomers [,] ( and ). Notably, both MucD(S217A) assemblies adopt a relatively closed cage conformation, as the largest surface pores are only 8 Å in the 12-mer and 18 Å in the 24-mer, compared with 20 Å and 50 Å in the DegP 12-mer and 24-mer, respectively (). Structural alignment of MucD and DegP 12-mers showed that the smaller pore in the MucD cage is mainly attributed to a shift of PDZ2* domains, where PDZ2* denotes a PDZ2 domain contributed by a neighboring trimer in inter-trimer interactions and oligomer assembly unless stated otherwise. This shift allows the “cage loops” (LC) to extend further toward the pore center (). Consistently, sequence alignment showed that the MucD LC is six amino acids longer than that of DegP (), allowing it to occupy more space and further restrict pore access. Superposition of the MucD 12-mer and 24-mer indicated that both assemblies use the same PDZ1-PDZ2*-mediated inter-trimer interaction, similar to what was previously reported for DegP 12-mer and 24-mer assemblies [] (). Thus, formation of distinct MucD oligomers appears to be enabled by conformational rearrangement of PDZ2 domains from neighboring trimers, facilitated by the flexible linker between the two PDZ domains within each protomer (). To gain insight into the interaction interface, we generated locally refined maps of constituent MucD trimers at 2.9 Å and 3.0 Å resolution from the 12-mer and 24-mer reconstructions, respectively. This was achieved by symmetry expansion followed by local refinement using a mask covering one trimeric subunit (, and ). In both maps, additional density was clearly visible near the protease and PDZ1 domains of each protomer (). Based on structural comparison with the DegP–hTRF1 (PDB: 8F0A) complex [] (), we assigned this density to the bound AlgK369-388 peptide. Structural alignment of the locally refined trimers from the 12-mer and 24-mer complexes showed that they are nearly identical, with a Cα RMSD of 0.43 Å (). This similarity indicates that the MucD trimer serves as the basic building block of higher-order oligomers and the oligomeric state does not substantially alter the substrate-binding mode (). Because the locally refined trimer from the 24-mer complex showed better overall map quality ( and ), we used this model for subsequent structural analysis unless otherwise stated. The fitted AlgK369-388 model revealed that the N-terminal 11 residues bind the protease domain of one protomer, whereas the last 4 residues of the C-terminus extend toward the PDZ1 domain of the adjacent protomer ( and ). Density was well resolved for most of the bound AlgK369-388 peptide, with consecutive side chains clearly distinguishable, despite the fact that the residues Q380-D384 could not be modeled due to the poor electron density ( and ). Notably, the MucD protease domain contains several conserved regulatory loops (LD, L1, L2, and L3) based on sequence and structure alignments with other bacterial HtrA-like proteases [,,] ( and ). Among these loops, LD, L1, and L3 regulate the switch between the active and inactive forms of the protease domain, while L2 determines the substrate specificity []. In the MucD structure, the protease domain adopts an open, catalytically competent conformation that accommodates the substrate peptide, as indicated by the displaced positions of these conserved regulatory loops that surround the bound peptide (). Alignment of MucD and DegP L2 loops revealed differences in the composition of substrate-contacting residues, which likely underlies their distinct substrate specificities ( and ). For example, R238 in MucD forms a salt bridge with D370 of AlgK369-388, which significantly increases the capacity of the L2 loop to engage negatively charged substrate residues. Furthermore, F236 in MucD forms hydrophobic contact with L372 of the bound AlgK369-388 and may also support π-π interactions with aromatic residues in other substrates (). Consistent with this role, substitution of F236 with alanine markedly reduced MucD proteolytical activity without disrupting oligomeric equilibrium: only limited AlgK degradation by MucD(F236A) was observed after 4 h of incubation (, , and ). These results indicate that F236-mediated hydrophobic interaction is important for substrate recognition. The side chain of the catalytic residue S217A points toward the peptide bond between substrate residues A375 and A376, defining the positions of the P1 and P1’ residues (). Analysis of the peptide–protease interface showed that the P1 side chain, A375, inserts into a shallow hydrophobic S1 pocket formed primarily by these residues I212, P214, and I235 in MucD ( and ). This arrangement is consistent with our MS analysis, which indicated a preference for small hydrophobic residues at the P1 position (). Indeed, in silico substitution of A375 with seven other hydrophobic residues showed that alanine, valine, isoleucine, and leucine can be accommodated, whereas methionine, phenylalanine, tyrosine, and tryptophan, which have larger side chains, would exceed the spatial constraints of the S1 pocket and cause steric clashes (). These structural observations agree well with the MS-derived specificity profile of the P1 site. At the C-terminus of the AlgK369-388 peptide, only the last 4 residues showed sufficient electron density for modeling, suggesting a relatively weak but defined interaction with PDZ1 (). The backbone carbonyl of substrate residue A388 forms hydrogen bonds with the backbone amides of MucD G273 and V274 (). In addition, substrate residue L387 makes hydrophobic contact with MucD V275 through side-chain interaction, further stabilizing peptide binding (). Notably, the side chain of A388 points into a hydrophobic pocket formed mainly by MucD residues L272, L331, P332, and V335 ( and ). Because the PDZ1 domain functions as a central regulatory module in HtrA-like proteases, linking substrate recognition to protease activation [], this hydrophobic PDZ1 pocket, together with adjacent residues V275 and L272, may recognize characteristic substrate C-terminal motifs and thereby promote allosteric activation of degradation. To test the role of PDZ1, we generated a MucD variant lacking this domain (MucD ΔPDZ1). This variant remained trimeric in solution (), consistent with the observation that trimer formation is primarily mediated by protease-domain interactions (). However, deletion of PDZ1 almost completely abolished proteolytic activity: no substantial degradation of AlgK or AlgX was detected after 4 h of incubation (), compared to the wild-type MucD ( and –). These results indicate that substrate engagement by PDZ1 is essential for efficient MucD-mediated proteolysis. The proteolytic activity of MucD is regulated via oligomerization Given that MucD exists in a dynamic equilibrium between trimeric and 12-mer states in solution, we next examined how oligomerization affects its proteolytic activity. Our structures show that higher-order MucD oligomers, including the 12-mer and 24-mer, are assembled through interactions between PDZ1 and PDZ2* domains (contributed by neighboring trimers) in an oligomerization mode reminiscent of DegP [] ( and ). In DegP, formation of large oligomers is stabilized by a strong hydrophobic interface involving PDZ1 residues L276, M280, and F289 and PDZ2* residues Y431, Y444, and L446 []. By contrast, the corresponding MucD interface is dominated by a relatively weak π-π interaction between F287 (PDZ1) and F467 (PDZ2*) (, right). Substitution of F467 with alanine yielded a variant that remained trimeric after purification, as shown by analytic SEC, and no longer underwent substrate-induced changes in oligomeric state (). These structure-guided mutagenesis results support a model in which PDZ1-PDZ2* interactions drive higher-order MucD assembly. Notably, MucD(F467A) degraded AlgK more efficiently than the wild-type MucD, accomplishing degradation of full-length AlgK within 1 h (). However, SDS-PAGE in combination with MALDI/TOF-MS analysis revealed accumulation of partially degraded AlgK fragments (, ), unlike wild-type MucD, which degraded AlgK completely without detectable large fragment bands ( and ). We propose that the MucD 12-mer adopts a closed cage conformation that restricts substrate access, thereby limiting its proteolytic activity against folded substrates ( and ). To test this model, we introduced cysteine substitutions at D282 and T469 to reinforce the PDZ1–PDZ2* interface, given that both residues are solvent-exposed at the interface and have side-chain lengths comparable to cysteine (). As MucD is a periplasmic protein with only one endogenous cysteine within its signal peptide, the nonreducing periplasmic environment should favor formation of specific disulfide bonds at the engineered sites []. Indeed, the purified MucD(D282C–T469C) mutant predominantly exists as a stable 12-mer, with no tendency to disassemble upon substrate addition (). SDS-PAGE analysis revealed prominent DTT-sensitive high-molecular-weight bands, indicating successful interdomain cross-linking via disulfide bonds (). If cross-linking occurred at every D282C-T469C pair across the PDZ1-PDZ2* interface within the 12-mer, nonreducing SDS-PAGE would be expected to show a band of ~160 kDa, corresponding to three disulfide-linked MucD molecules from adjacent trimers (). Of note, additional bands near 110 kDa likely reflect incomplete cross-linking across the interfaces, whereas minor bands above 250 kDa may represent nonspecific aggregation caused by random cross-linking in vivo (). Functionally, MucD(D282C–T469C) degraded AlgK much more slowly than wild-type MucD ( and ), with both full-length and partially degraded AlgK still detectable after 4 h of reaction (), supporting the hypothesis that the 12-mer is less proteolytically active. Although cage assembly does not fully abolish proteolysis, the dynamic equilibrium between the 12-mer and trimer likely enables complete substrate degradation while preventing accumulation of partially degraded fragments, thereby supporting periplasmic protein quality control. Within each MucD trimer, protomer association is stabilized by extensive hydrophobic contacts and hydrogen bonds between protease-domain residues at the interface (), similar to the trimeric interface observed in P. syringae MucD (psMucD; PDB: 8K2Y) []. At the N-terminal region of MucD, where the three protomers converge, F30 and W171 form a triangular aromatic arrangement stabilized by edge-to-face or stacked π-π interactions (). Additional π-π interactions occur between F236 from one protomer and F179–F181 from a neighboring protomer (). Hydrogen bonds between the amide groups of Q207 and R195 from one protomer and the hydroxyl group of S184 from another protomer further reinforce the trimer stability (). Structural alignment showed that trimeric psMucD (PDB: 8K2Y) and P. aeruginosa MucD (paMucD; PDB: 22SK, from the 24-mer reconstruction) are highly similar, with a Cα RMSD of 0.69 Å (). In both structures, the major regulatory loops of the protease domain, including L1, L2, L3, and LD, adopt open conformations that allow substrate binding ( and ). These observations support the view that trimeric MucD represents an active state for proteolysis. MucD represses alginate production by directly degrading alginate synthesis proteins MucD has been reported as a negative regulator of alginate production in P. aeruginosa, as the deletion of mucD leads to alginate overproduction in the wild-type laboratory PAO1 strain [,]. To determine whether this regulation depends on MucD protease activity, we measured alginate-producing levels in ΔmucD mutants in the PAO1 background using the carbazole assay, a quantitative method for detecting alginate in P. aeruginosa [,]. Notably, alginate production was below the detection limit in wild-type PAO1 but increased significantly in the ΔmucD mutants (), in line with the previous reports [,]. Complementation with wild-type mucD or proteolytically active MucD(F467A) variant expressed from the commonly used plasmid pUCP24 [] reduced alginate production to an undetectable level (). Growth monitoring over 24 h confirmed that these differences were not caused by changes in cell fitness (). In contrast, complementation with either the proteolytically inactive MucD(S217A) variant or the disulfide-locked 12-mer MucD(D282C-T469C) variant did not significantly reduce alginate production (). These results indicate that MucD-mediated repression of alginate production requires proteolytic activity and is impaired when MucD is trapped in the resting 12-mer state. A previous study reported that deletion of mucD in PAO1 increases promoter activity of the alg operon []. To assess whether the elevated alginate production observed here was accompanied by transcriptional changes in algK and algX, quantitative reverse transcription PCR (RT-qPCR) for the wild-type, ΔmucD, and ΔmucD::mucD strains was carried out. The results showed that the mRNA levels of algK and algX did not differ significantly among the strains. By contrast, algD, the proximal gene in the operon encoding the rate-limiting enzyme of the pathway, increased ~2.8-fold in the ΔmucD strain and restored to wild-type levels upon mucD complementation (). The different behavior of the proximal algD transcript and the distal algK and algX transcripts likely reflects natural transcriptional attenuation or differential mRNA stability across this large 12-gene biosynthetic cluster, consistent with previous evidence for increased promoter activity in ΔmucD cells [,,]. Given that AlgK is a MucD substrate in vitro, we next examined the AlgK protein level in vivo by western blot in PAO1, ΔmucD, and ΔmucD::mucD strains. Our results revealed that AlgK protein level was markedly increased in the ΔmucD mutant and restored by mucD complementation (). Together with its degradation by MucD in vitro ( and ) and its reported association with MucD in vivo [], these findings support AlgK as a physiological substrate of MucD. AlgX can also be degraded by MucD in vitro ( and ) and was previously identified in the same periplasmic complex, however, its substrate status in vivo requires more experiments with complementary approaches. Previous studies demonstrated that alginate export and modification require a large complex composed of AlgK, AlgX, AlgJ, and AlgF in P. aeruginosa []. To test whether MucD also targets other alginate synthesis proteins in this complex, we purified soluble constructs of AlgJ79-379 and AlgF30-216 [,] and examined their susceptibility to MucD-mediated degradation. Interestingly, AlgF was degraded by MucD, whereas AlgJ was resistant ( and ), suggesting that post-translational regulation by MucD is selective and restricted to particular components of the alginate biosynthetic machinery. Collectively, these results suggest that MucD can alternatively regulate alginate production by directly degrading alginate biosynthetic proteins. Discussion MucD is an HtrA-like serine protease that plays a critical role in stress adaptation, virulence, and regulation of alginate biosynthesis in P. aeruginosa [,]. In this study, we newly identified AlgK and AlgX as MucD substrates. Both proteins are essential components of the alginate biosynthetic machinery, responsible for alginate chain export and acetylation, respectively. By combining LC-MS with in silico mapping of the proteolytic fragments using UMSAP, we defined MucD cleavage preferences and sequence selectivity. Notably, our cryo-EM structures of the MucD 12-mer and 24-mer complexed with a substrate peptide uncover the structural basis for MucD oligomerization, protease-substrate interactions, and specificity, which nicely rationalizes our biochemical characterization data. Mutagenesis of interface residues demonstrated that the proteolytic activity of MucD is modulated by its oligomeric state. Finally, we showed that beyond the known transcriptional-level regulation, MucD alternatively represses alginate production in vivo through degradation of alginate biosynthesis proteins. These findings establish a post-translational regulatory mechanism for alginate biosynthesis and provide mechanistic insight into the function of bacterial HtrA-like proteases. Oligomerization regulates MucD proteolytic activity Oligomerization is widely employed for HtrA-like proteases to regulate their activity []. DegP, for example, forms a proteolytically inactive hexamer through the interaction between the extended LA loops of two “face-to-face” trimers []. Upon substrate binding, DegP rapidly converts into higher-order oligomers (12-mer and 24-mer), driven by hydrophobic interactions between PDZ1 and PDZ2* domains, and then returns to the resting state after substrate degradation []. These cage-like assemblies have also been proposed to support chaperone activity by assisting the folding of outer membrane proteins []. However, their physiological significance remains debated, as higher-order oligomers are not strictly required for proteolysis or high-temperature survival []. In contrast, MucD follows a distinct regulatory pattern, namely maintaining an equilibrium between trimeric and 12-mer states in solution ( and and ). Substrate binding promotes the formation of the trimer and disassembly of the 12-mer, suggesting that the trimer is the preferred active form ( and ). Mutagenesis of inter-trimer interface residues further supports the notion that the 12-mer represents a resting or less active state, while the trimer mediates efficient proteolysis (). Nevertheless, none of the MucD variants tested fully reproduced wild-type substrate turnover, and both oligomer-biased mutants accumulated partially cleaved AlgK fragments ( and ). These observations suggest that dynamic exchange between the 12-mer and trimer may be important for processive degradation, perhaps by facilitating substrate remodeling or unfolding to expose buried cleavage sites. The distinct oligomerization-driven regulation of MucD and DegP is structurally relevant. While both enzymes use a similar trimeric building block, MucD has a shorter LA loop than DegP, which may limit its ability to form a stable hexamer (). Furthermore, the PDZ1-PDZ2* interface that sustains the MucD 12-mer or 24-mer cage is mediated mainly by a π-π interaction between F287 and F467 (). This interface is substantially weaker than the multi-residue hydrophobic interface in DegP, where PDZ1 residues L276, M280, and F289 interact with PDZ2* residues Y431, Y444, and L446 []. These differences likely explain why MucD and DegP show distinct patterns of cage assembly and disassembly during proteolysis ( and ). Interestingly, substrate addition to the proteolytically inactive MucD(S217A) mutant shifted the equilibrium toward 12-mer and 24-mer assemblies, a behavior distinct from that of wild-type MucD (, , and ). This finding raises the possibility that inhibitory peptides or small proteins, analogous to naturally produced serine protease inhibitors such as serpins [], could regulate MucD by stabilizing the higher-order oligomeric states as post-cleavage intermediates. Ecotin, a broad-spectrum periplasmic serine protease inhibitor found in many bacteria, including E. coli and P. aeruginosa, provides one example of this type of post-translational regulation []. Although we did not observe 24-mer formation for wild-type MucD, the 12-mer and 24-mer cage assemblies may also relate to a potential chaperone-like function. Under stress conditions, membrane or periplasmic proteins with critical function may be recognized by MucD and trigger the cage formation. Since the higher oligomers of MucD adopt a more closed formation (), they may shield and stabilize captured substrates from the disordered periplasm environment. However, we did not detect proteins co-purifying with MucD, and no putative chaperone partners have yet been reported. Given that the chaperone activity is highly condition-dependent, whether MucD has chaperone activity remains an open question. Substrate recognition and specificity Hydrophobic residues are preferred cleavage sites for many HtrA-like proteases [,]. Consistent with this feature, structural analysis of the MucD S1 pocket together with MS-based mapping of cleavage products demonstrated that MucD preferentially cleaves at hydrophobic residues ( and ). Substrate recognition, however, is not determined by the protease domain alone. Our structure of the MucD-AlgK369-388 complex revealed that PDZ1 also engages the substrate C-terminus through backbone hydrogen bonds and hydrophobic contacts, particularly with the side chain of the C-terminal end A388 positioned in a hydrophobic pocket analogous to the S1 site ( and ). In the complex structure ( and ), the 13-residue peptide spanning from the P1’ residue A376 to the C-terminal end residue A388 closely matches the average length of AlgK and AlgX degradation fragments, which are 13 and 14 residues, respectively (). Together with the observation that the deletion of the PDZ1 domain almost completely abolishes proteolytic activity, these findings support a model in which PDZ1 acts as a receptor for the substrate C‐terminus following an initial cleavage event. This interaction may stabilize the protease-substrate complex and help define fragment length during processive degradation. Regulation of alginate biosynthesis and physiological implications MucD has long been recognized as a negative regulator of alginate biosynthesis. Previous studies have suggested a transcriptional regulation function for MucD, as deletion of mucD increases promoter activity of the alg operon [], consistent with our observation that algD mRNA levels were significantly elevated in the ΔmucD strain (). Our results further demonstrate that MucD proteolytic activity is required for this regulatory function in wild-type P. aeruginosa. Notably, transcript levels of the essential alginate biosynthesis genes, algK and algX, remained unchanged in ΔmucD cells, whereas AlgK protein abundance increased significantly ( and ). In large bacterial operons such as the alg operon, polycistronic transcripts can undergo progressive downstream attenuation, internal termination, or segmented mRNA decay, uncoupling steady-state transcript levels of distal structural genes (algK and algX) from primary promoter activity []. Collectively, these findings indicate that MucD regulates alginate biosynthesis through both negative transcriptional control of the alg operon and post-translational quality control of key assembly components, including AlgK and AlgX, as a model is proposed to demonstrate MucD oligomerization assembly, its activity regulation (), and its involvement within the alginate biosynthesis multiprotein complex (). MucD is a periplasmic HtrA-like serine protease that involves pathways critical for bacterial survival and virulence, particularly for P. aeruginosa, an opportunistic pathogen capable of infecting plants, animals, and even humans. Loss of mucD reduces virulence in several infection models [,], underscoring its physiological importance. While our study elucidates a role for MucD in post-translational control of alginate biosynthesis, its broader involvement in other processes—such as host-pathogen interactions, antibiotic resistance, and biofilm formation—remains to be explored. Methods Bacteria strains and plasmids All recombinant plasmids were generated by Gibson assembly using the ClonExpress II One Step Cloning Kit (Vazyme). The mucD gene variants encoding MucD S217A, F467A and D282C-T469C mutants were generated by site-directed mutagenesis. Seamless deletion of mucD gene in PAO1 strain was conducted using the pEX18Tc plasmid. Details of bacterial strains and plasmids used in this study are listed in . Protein purification AlgK, AlgX as well as MucD and its variants from P. aeruginosa were purified similarly. Briefly, the plasmid pMMB67EH containing the relevant gene with a hexa-histidine (6× His) tag at the C-terminus was transformed into the P. aeruginosa PAO1 strain by electroporation for protein expression. Bacteria were cultured at 37 °C in Luria-Bertani (LB) broth (10 g/L tryptone, 5 g/L yeast extract, and 5 g/L NaCl, pH 7.0–7.5) supplied with 150 μg/mL carbenicillin to OD600nm reached 0.8 with agitation at 200 rpm. Proteins expression was induced by 0.8 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and bacteria were grown at 16 °C for 16 h with agitation at 160 rpm. Bacteria were harvested by centrifugation (4,000 g, 20 min) and resuspended in Buffer A (50 mM HEPES and 500 mM NaCl, pH 7.5). Cells were lysed by sonication and then centrifuged at 20,000g for 1 h to remove insoluble fractions. The supernatant was collected and incubated with nickel–nitrilotriacetic acid (NTA) beads (Cytiva), pre-equilibrated with Buffer A, at 4 °C for 1 h with gentle rotation. The proteins were purified by gravity flow chromatography by extensive washing with Buffer E (20 mM HEPES and 150 mM NaCl, pH 7.5) supplemented with 30 mM imidazole and subsequent elution with Buffer E supplemented with 300 mM imidazole. Eluted proteins were concentrated using a 30-kDa molecular weight cutoff (MWCO) concentrator (Millipore) and loaded onto a HiLoad 16/600 Superdex 200 column (Cytiva) pre-equilibrated with Buffer E. The fractions containing the target proteins were pooled, concentrated, and stored by snap freezing for downstream experiments. AlgJ79-391 and AlgF30-216 were cloned from the P. aeruginosa genome into the pDBHT-2 vector with a 6 × His tag at the N-terminus, separately. The recombinant vectors were transformed into the E. coli BL21(DE3) strain for protein expression and purification as described previously [,]. Purified AlgJ79-391 and AlgF30-216 proteins were buffer exchanged to Buffer E, concentrated, and stored by snap freezing for downstream experiments. In vitro protein degradation assays Eighty μM MucD and 80 μM substrates (AlgK, AlgX, AlgKX, AlgJ79-391 and AlgF30-216) were mixed in Buffer E and incubated at 37 °C. At each time point, an aliquot was taken and diluted 400-fold with Buffer E. 50 ng of samples supplied with NuPAGE LDS Sample Buffer (Invitrogen) were incubated at 70 °C for 10 min before loading onto a 4%–12% NuPAGE Bis-Tris Mini Protein Gel (Invitrogen). Gel electrophoresis was conducted in MES buffer at 200V for 38 min. After electrophoresis, the gel was stained by SYPRO Ruby (Thermo-Fisher Scientific) and visualized by the ChemiDoc Touch Image System (Bio-Rad). Degradation assays of MucD variants were conducted following a similar process. Analytical size exclusion chromatography (SEC) The purified 12-mer and trimer population of wild-type MucD and MucD(S217A) at a concentration of 120 μM were loaded to an analytical SEC column (Superose 6 Increase 10/300 GL, Cytiva). To analyze the proteolysis process or form complex, MucD and its variants were incubated with AlgK at a 1:1 molar ratio (120 μM) at 37 °C for 15 min or 2 h prior to application to the analytical SEC column (Superose 6 Increase 10/300 GL, Cytiva). Apparent molecular weights (MW) of different MucD oligomeric states were calculated by interpolating from MW versus elution volume standard curve, which was generated by using a Gel Filtration Standard (Bio-Rad) as per the manufacturer’s guidelines. Analysis of MucD proteolysis products by LC-MS For proteolysis, 80 μM MucD was incubated with 80 μM AlgK or AlgX in Buffer E at 37 °C for 8 h. Samples were then centrifuged using a 10-kDa molecular weight cutoff (MWCO) concentrator (Millipore) to remove intact proteins. Peptides resulting from AlgK and AlgX degradation in the flowthrough were subjected to LC-MS analysis. The mass spectrometry experiments were performed in the Mass Spectrometry Core Facility in the School of Biological Sciences, Nanyang Technological University. The fractionated peptides were separated and analyzed using a Vanquish Neo UHPLC System coupled to an Orbitrap Exploris 480 (ThermoFisher Scientific). Separation was performed on an EASY-Spray 75 μm × 15 cm column packed with PepMap Neo C18 2 μm, 100 Å (ThermoFisher Scientific) using solvent A (0.1% formic acid) and solvent B (0.1% formic acid in 80% ACN) at a flow rate of 300 nL/min with a 60 min gradient. Peptides were then analyzed on an Orbitrap Exploris 480 apparatus with an EASY nanospray source (ThermoFisher Scientific) at an electrospray potential of 2.0 kV. A full MS scan (200–2,000 m/z range) was acquired at a resolution of 60,000. Dynamic exclusion was set as 30 s. The resolution of the higher energy collisional dissociation (HCD) spectra was set to 30,000. The automatic gain control (AGC) settings of the full MS scan and the MS2 scan were 200% normalized AGC target and standard, respectively. The data-dependent mode was cycle time. The time between each master scan was 2 s. An isolation window of 1.6 m/z was used for MS2. Single and unassigned charged ions were excluded from MS/MS. For HCD, the normalized collision energy was set to 32%. Raw data files were processed and searched using Proteome Discoverer 2.2 (ThermoFisher Scientific). The raw LC-MS/MS data files were loaded into Spectrum Files (default parameters set in Spectrum Selector). The Mascot algorithm was then used for data searching to identify proteins using the following parameters: missed cleavage of three; dynamic modifications were oxidation (+15.995 Da) (M), and deamidation (+0.984 Da) (N and Q), acetylation (+42.0105 Da) (K), methylation (14.0157 Da) (K and R), (28.0313 Da) (K and R) and phosphorylation (+79.966 Da) (S, T, and Y). Percolator was applied to filter out the false MS2 assignments at a strict false discovery rate of 1% and a relaxed false discovery rate of 5%. Analysis of detected peptides by UMSAP Identified peptides were further analyzed using the Targeted Proteolysis module of UMSAP 2.3.3 []. The significance level was set to 0.05 and the minimum score value to 10. A Log2 transformation was applied to the data before the analysis. The amino acid distribution around the cleavage sites included five residues in each direction. Calculation of the relative frequency of cuts by UMSAP followed a similar procedure as described in the reference []. Multiple and pairwise sequence alignment Protein sequences were retracted from UniProt []. The UniProt entry identifiers for the sequences are as follows: paMucD (G3XD20); psMucD (Q9AQD1); ecDegP (P0C0V0); ecDegQ (P39099). For both multiple and pairwise sequence alignment, sequences were input into Snapgene software. Cryo-EM samples preparation and data collection The AlgK369-388 peptide (synthesized by GeneScript) was dissolved in 100% DMSO at the concentration of 2.4 mM. To form the MucD(S217A)-AlgK369-388 complex, AlgK369-388 was diluted by Buffer S (20 mM HEPES and 50 mM NaCl, pH 7.5) to 120 μM (0.5% DMSO) and incubated with 24 μM MucD(S217A) in Buffer S at 37 °C for 16 h. The sample was loaded onto a Superose 6 Increase 10/300 gel filtration column (Cytiva) pre-equilibrated with Buffer S. The fractions containing the 12-mer or 24-mer population were pooled and concentrated to 5.0 mg/mL. For both complexes, a total of 4 μl of protein sample at a concentration of 5.0 mg/ml was applied to the Quantifoil Cu R1.2/1.3 grids. The grids were blotted using a Vitrobot Mark IV (Thermo-Fisher Scientific) operated at 4 °C and 100% humidity with a blotting time of 1 s, blotting force of 1, and waiting time of 1 s. Datasets were collected on a Titan Krios (Thermo-Fisher Scientific) 300-kV electron microscope equipped with a GIF Quantum energy filter and a Falcon4i detector with a defocus range between −0.6 μm and −1.2 μm. The magnification was ×130,000 with a pixel size of 0.97 Å. Cryo-EM data processing Micrographs were imported to Cryosparc4.1 for patch motion correction and patch CTF estimation []. Micrographs with an estimated CTF resolution higher than 6 Å were selected for subsequent processing. Particles were auto-picked by the blob picker, manually curated, and extracted with a box size of 360 and a pixel size of 3.04 Å. All extracted particles were subjected to a reference-free 2D classification. Good classes with clear particle features were selected and used for initial model building and heterogeneous refinement. Only the reasonable models were selected, and the good particles were re-extracted with a box size of 320 and a pixel size of 1.216 Å. The re-extracted particles were used for homogeneous refinement and, subsequently, local refinement. The final resolutions of all resulting maps were calculated based on gold standard Fourier shell correlation (FSC) = 0.143. AlphaFold–predicted models of MucD (database accession code: AF-G3XD20-F1) were used for model building [], which were manually adjusted with COOT [], and real-space refinement was done in Phenix Suite []. All the structure-related figures were generated by ChimeraX []. Alginate production, purification, and detection Alginate production and purification were carried out as described previously [] with several changes. Briefly, an overnight bacterial culture was diluted in 10 mL of alginate-producing medium (100 mM monosodium glutamate, 7.5 mM NaH2PO4, 16.8 mM K2HPO4, and 10 mM MgSO4, pH = 7.0) to an OD600nm of 0.04 and incubated at 37 °C for 24 h with agitation at 200 rpm. For culturing mucD complementary strains, the medium was supplemented with gentamicin at a final concentration of 50 μg/mL. Cells were removed by centrifugation (4,000 g, 15 min) and the supernatant containing alginate was collected. To precipitate alginate, three volumes of ice-cold absolute ethanol was added to the supernatant and incubated at 4 °C for 2 h with gentle rotation. Precipitated alginate was collected by centrifugation at 4 °C (4,000 g, 20 min) and washed with 10 mL ice-cold absolute ethanol twice. The excess ethanol was removed by air drying samples at 25 °C for 6 h. 5 mL of distilled water was added to dissolve alginate by thorough vortex and subsequent incubation at 42 °C for 30 min. Alginate was quantitatively detected using the carbazole assay for uronic acids as described previously [,] with several modifications. Briefly, a borate-sulfuric acid solution (0.1 M H3BO3 in concentrated H2SO4) and a carbazole reagent (0.1% (w/v) of carbazole in absolute ethanol) were prepared. 3 mL of the borate-sulfuric acid solution was chilled on ice. A 500 μL aliquot of alginate solution was slowly added to the top of the acid mixture. After a brief vortex at low amplitude, 100 μL of the carbazole reagent was added to the acid sample mixture on ice. Following a brief vortex, samples were incubated at 55 °C for 30 min and subsequently cooled at room temperature for 5 min. The absorbance was measured at 530 nm using a spectrophotometer with three biological repeats. RNA extraction, reverse transcription, and quantitative PCR The overnight culture was diluted in 3 mL of alginate-producing medium to an initial OD600nm of 0.04 and incubated at 37 °C for 24 h with agitation at 200 rpm. After harvesting by centrifugation, bacteria were normalized to an OD600nm of 1.0 in 1 mL PBS buffer. Total RNA was isolated using a PureLink RNA Mini Kit (Invitrogen) and RNA concentrations were determined using a NanoDrop spectrophotometer (Thermo-Fisher Scientific). Reverse transcription and cDNA synthesis were carried out by the Maxima First Strand cDNA Synthesis Kit (Thermo-Fisher Scientific). Quantification was performed with the SYBR Select Master Mix (Applied Biosystems) on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad). The small ribosomal subunit protein gene rpsL was used as an internal control. Antibody production and western blot Purified AlgK protein was used as the antigen to generate rabbit polyclonal antiserum using a standard immunization protocol. The anti-AlgK antiserum was further purified by Protein A affinity chromatography (performed by RHBiolabs). A monoclonal rabbit antibody against the bacterial RNA polymerase β-subunit was purchased from Abcam. An overnight bacterial culture was diluted in 3 mL of alginate-producing medium to an OD600nm of 0.04 and incubated at 37 °C for 24 h with agitation at 200 rpm. Cells were harvested by centrifugation and normalized to a final OD600nm of 1.0 in 1 mL PBS buffer. The cell pellet from the normalized resuspension was collected by centrifugation and resuspended in 100 μL of lysis buffer (PBS buffer, 0.1% (w/v) SDS, and 1% (w/v) Triton X-100). Three freeze-thaw cycles using liquid nitrogen were applied to lyse bacteria thoroughly. Cell lysate was then centrifuged at 15,000 g for 10 min, and the supernatant was combined with SDS loading dye. Samples were boiled at 95 °C for 10 min prior to loading the clarified lysate onto a 10% (v/v) Tris-Glycine polyacrylamide gel. Following SDS-PAGE, proteins were transferred to a PVDF membrane for immunoblotting. The membrane was blocked using a blocking buffer (5% (w/v) skim milk dissolved in TBST buffer) at 4 °C overnight. Blots were washed twice with TBST buffer and the membrane was cut along the 100 kDa molecular weight marker. The membrane containing proteins less than 100 kDa was incubated with anti-AlgK antibody (produced by RHBiolabs) at a 1:2000 dilution in the blocking buffer, while the membrane containing proteins larger than 100 kDa was incubated with anti-RNA polymerase β-subunit antibody (Abcam) at a 1:5000 dilution in the blocking buffer. Both membranes were incubated with primary antibodies at 25 °C for 2 h and then washed three times with TBST. The primary antibodies were probed with a goat anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibody (Abcam) at 1:5000 dilution in the blocking buffer for 2 h at 25 °C. Blots were washed three times with TBST. AlgK and RNA polymerase β-subunit bands were detected using the SuperPico ECL Chemiluminescence Kit (Vazyme). Blots were imaged using the ChemiDoc Touch Image System (Bio-Rad). Statistical analysis All raw image data were processed with ImageJ software. Normally distributed data were presented as mean ± SD. Comparisons among three or more groups were performed using one-way ANOVA followed by Tukey’s post hoc test. A p-value of less than 0.05 was considered statistically significant. The statistical parameters were shown in the figures and described in the figure legends. All statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, USA).