Engineering the stambomycin modular polyketide synthase yields 37-membered mini-stambomycins

Design of engineering experiments based on classical modular boundaries

The stambomycin PKS comprises 25 modules distributed among 9 polypeptides (Pks1–9)³³ (Fig. 1a) (Note: throughout the text, the stambomycin genes have been numbered in accordance with ref. ³³). Module 12 of Pks4 notably houses a broad-specificity AT domain which gives rise to the six characterized stambomycin family members (A–F), which differ in the alkyl functionality at the C-26 position^33,34. To access abridged derivatives using the classical module boundaries, we reasoned that we could engineer intersubunit interfaces by suitable manipulation of docking domains. Encouragingly, the extreme C- and N-termini of all subunits (with the exception of the N-terminus of Pks1 and the C-terminus of Pks9) contain sequences with convincing homology to previously identified DDs^26,35 (the C-terminal DDs are referred to hereafter as ^CDDs and their partner N-terminal DDs as ^NDDs). By bioinformatics analysis, we were able to confidently assign the DDs acting at 6 of the 8 interfaces to the type 1a class²⁶, and the remaining two sets of DDs as type 1b³⁵ (Supplementary Fig. 1). In both cases, docking occurs between an α-helical ^CDD and a coiled-coil formed by the ^NDD, with specificity achieved via strategically placed charge:charge interactions at the complex interface (Supplementary Fig. 1)^26,35.

Among the type 1a junctions, there were notably two sets which appeared compatible in terms of the translocated substrate: Pks 3/4 + 7/8 and Pks 4/5 + 8/9 (Supplementary Fig. 2). Specifically, the functional groups at the critical α- and β-positions^17,36 of the transferred chains are identical at these junctions, and correspondingly, the downstream KSs show similarities across several sequence motifs previously correlated with substrate specificity^17,24,37 (Supplementary Fig. 2). Targeting such interfaces thus allowed us, at least in principle, to overcome the functional block to the engineered systems represented by poor recognition of the incoming substrate by the directly downstream KS domain²⁵. Ultimately, we aimed to create an interface between Pks subunits 4 and 9 for two principal reasons. First, as mentioned earlier, Pks4 is at the origin of the structural variation between the stambomycin family members, and thus we anticipated that maintaining the subunit within the hybrid system would give rise to a corresponding series of truncated analogues, providing important evidence for their identities. Second, it was genetically more practical to modify the second set of interfaces due to splitting of the genes encoding the PKS subunits between two loci (Fig. 1 and Supplementary Fig. 4).

To establish the Pks4/Pks9 junction, we initially modified the ^CDD of Pks4 (^CDD₄) to match that of Pks8 (the natural partner of the ^NDD of Pks9 (^NDD₉)), either by site-directed mutagenesis (SDM) of residues previously identified as key mediators of interaction specificity (construct ^CDD₄ SDM; Supplementary Fig. 3 and Supplementary Table 1)²⁶, or by exchange of the complete ^CDD docking α-helix of ^CDD₄ for that of ^CDD₈ (construct ^CDD₄ helix swap; Supplementary Fig. 3 and Supplementary Table 1)³⁸. Modifying the ^CDD₄ specificity code to match that of ^CDD₈ required mutation of 3 residues, while for the ^CDD₄ helix swap, the terminal 16 amino acids of ^CDD₄ were exchanged for the corresponding 15 residues of ^CDD₈ (Supplementary Fig. 3 and Supplementary Table 1). The genetic alterations were carried out in two distinct PKS contexts: (i) in the presence of the intervening subunits 5–8, which allowed for the possibility of competitive interactions between modified Pks4 and both Pks5 and Pks9; and (ii) removing the intervening subunits 5–8, thus eliminating competition for binding of Pks4 by Pks5, and of Pks9 by Pks8 (Supplementary Fig. 3). We further generated a mutant in which Pks subunits 5–8 were deleted but no modification was made to ^CDD₄, in order to judge the intrinsic capacity of Pks4 and Pks9 to interact. Furthermore, genetic engineering was carried out in parallel by both PCR-targeting³⁹ and CRISPR-Cas9⁴⁰ (Supplementary Figs. 5 and 6), in order to directly compare the efficacy of these two approaches, as well as evaluate the effect of the short scar sequence remaining in the chromosome following PCR-targeting.

The ^CDD₄ SDM and ^CDD₄ helix swap sequences were introduced in parallel into the S. ambofaciens genome. As previous work has shown that production from the stambomycin biosynthetic gene cluster requires activation by constitutive overexpression of a pathway-specific LAL (Large ATP-binding regulators of the LuxR family) regulator³³, we additionally introduced the LAL overexpression plasmid (pOE484) into each of the mutants, using the empty parental plasmid (pIB139⁴¹) as a control. In total, this strategy resulted in 20 targeted strains harboring interface mutants (where K7N refers to PCR-targeting and CPN to CRISPR-Cas9 engineering): K7N1/pIB139, K7N1/OE484, K7N2/pIB139, K7N2/OE484, K7N3/pIB139, K7N3/OE484, K7N4/pIB139, K7N4/OE484, K7N5/pIB139, K7N5/OE484, K7N6/pIB139, K7N6/OE484, CPN1/pIB139, CPN1/OE484, CPN2/pIB139, CPN2/OE484, CPN4/pIB139, CPN4/OE484, CPN5/pIB139, CPN5/OE484 (Table 1, Supplementary Data 1–3; despite extensive efforts the CPN3 mutant strain was not obtained). The principal difference between the K7N and CPN series of constructs is the presence of a 33 bp scar sequence between the modified pks4 and pks9 genes (Supplementary Fig. 4). Construct K7N6 was assembled specifically to test the effect of this region, without any further modification to ^CDD₄ and the intervening pks5–pks8 genes.

With the exception of K7N3, CPN4, and CPN5, extracts of the engineered mutant strains harboring pOE484 were analyzed by high performance liquid chromatography heated electrospray ionization high-resolution mass spectrometry (HPLC-ESI-HRMS) on a Dionex UItiMate 3000 HPLC coupled to a Q Exactive^TM Hybrid Quadrupole-Orbitrap^TM Mass Spectrometer, and compared to extracts of the control strain containing pIB139⁴¹ as well as the wild-type S. ambofaciens, using SIEVE 2.0 screening software. All extracts were subsequently analyzed on a Thermo Scientific Orbitrap LTQXL and/or an Orbitrap ID-X Tribrid Mass Spectrometer, and the data inspected manually. Metabolites not present in the control strains are listed in Table 1 and Supplementary Table 2.

It must be noted that the low yields of the target mini-stambomycins (vide infra) precluded their purification, and consequently full structure elucidation by NMR and use as quantification standards. However, we were able to obtain convincing evidence for their identities based on exact masses obtained from high-resolution mass spectra, and detailed comparative analysis of MS and MS² data with those acquired on a series of shorter derivatives which were produced in substantially higher amounts (as detailed in the respective Supplementary figures). Further support for the identity of multiple metabolites was afforded by genetic engineering controls. In the absence of authentic standards for the stambomycin derivatives, we evaluated the use of two surrogates for quantification: the parental stambomycins 1A/B³³ (Supplementary Fig. 7–9), and linear, 50-deoxy derivatives of stambomycins 2A/B (Supplementary Figs. 10 and 11). The latter compounds were purified from a previously described strain in which the C-50 hydroxylase SamR0479 (Fig. 1a) had been inactivated⁴². This analysis notably revealed that the detection sensitivity towards the 50-deoxystambomycins 2 whether using MS or UV, was dramatically lower than for the parental stambomycins 1 (Supplementary Fig. 11). In the case of the MS analysis, we can clearly attribute this difference to the presence in 1 of β-D-mycaminose which contains an easily protonatable nitrogen, as it is absent in 2. Indeed, analysis of erythromycin A 3 which contains an alternative amino sugar, β-D-desosamine, showed it to be detected with similar sensitivity to 1 (Supplementary Fig. 12). Thus, overall, to permit an estimation of yield ranges for the engineered metabolites, we generated a standard curve based on stambomycins 1A/B for which we could detect a 25,000-fold range of concentrations (0.00001–0.25 mg mL⁻¹). Using this curve directly then provided the lower yield limit for the derivatives, while introduction of a correction factor (×206) based on the 50-deoxystambomycins 2, furnished the upper yield limit. Importantly, the maximum yields calculated directly from a limited calibration curve based on 2, did not differ substantially from those determined using the correction factor (Supplementary Table 3).

The first result is that the K7N6/OE484 mutant yielded a similar metabolic profile to S. ambofaciens wt (22 ± 3 mg L⁻¹ of stambomycins 1 (Supplementary Table 4), 73% relative yield), showing that the scar sequence impacted stambomycin production, but not dramatically (Fig. 3). By contrast, no stambomycins were observed, as anticipated, in all constructs in which Pks5–Pks8 had been removed (K7N1−3; CPN1, 2) (Fig. 3). Stambomycins 1 were present, however, in strains K7N4 and CPN4 harboring ^CDD₄ site-directed mutations and in the ^CDD₄ helix swap strain CPN5, all of which still contained Pks5–Pks8, albeit at reduced amounts relative to the wild-type (18, 23, and 14% of wt, respectively) (Fig. 3 and Supplementary Table 4). (Surprisingly, the metabolic profile of K7N5 reproducibly differed from that of CPN5, as no stambomycin-related metabolites were detected from K7N5 (Fig. 3)). These data suggested that while the mutations introduced into ^CDD₄ negatively impacted the interaction with ^NDD₅, they were not sufficient to disrupt natural chain transfer between Pks4 and Pks5. Thus, DD engineering to alter partner choice should be accompanied by removal of competing intersubunit interactions.

We did not find any evidence in the DD engineering experiments for any of the target 37-membered metabolites (Supplementary Figs. 3 and 13). However, all strains in which stambomycin production was abolished (Table 1) exhibited four peaks in common (Fig. 3b and Supplementary Fig. 13) (peaks potentially corresponding to additional derivatives were observed, but none were shared between multiple strains). The determined exact masses and MS/MS analysis (as exemplified by strain CPN2/OE484, Fig. 3b) correspond to truncated derivatives of stambomycins A/B and C/D respectively, following premature release from modules 12 and 13 of Pks4 (compounds 4–7, Fig. 3d and Supplementary Figs. 13–15; ca. 8-fold greater yield of the module 13 products (Supplementary Table 5)). Further support for the identity of these shunt compounds was obtained by grafting the chain-terminating (type I) thioesterase (TE) domain from the C-terminal end of Pks9 to the C-terminus of Pks4 in order to force chain release at this stage. Indeed, identical compounds were produced, but at 17-fold increased yield relative to CPN2/OE484, consistent with active off-loading of the chains (Fig. 3c, Supplementary Figs. 16–18, and Supplementary Table 5).

Based on their exact masses, both sets of shunt metabolites were hydroxylated on a single carbon, while none were found to bear the β-mycaminose of the mature stambomycins, consistent with the absence of the tetrahydropyran moiety to which it is normally tethered. To determine the location of the hydroxylation and therefore the hydroxylase responsible, we inactivated in mutant CPN2/OE484 the genes samR0478 and samR0479 encoding respectively, the stambomycin C-28 and C-50 cytochrome P450 hydroxylases (Fig. 1a)⁴². While extracts of CPN2/OE484/Δ478 were unchanged relative to CPN2/OE484 (i.e., the hydroxyl group was still present), the CPN2/OE484/Δ479 mutant exhibited four peaks with masses and fragmentation patterns corresponding to the deoxy shunt products (Fig. 3, Supplementary Figs. 19–21 (compounds 8–11) and Supplementary Table 5). Taken together, these data show that the unusual online modification catalyzed by SamR0479⁴², which is necessary for macrocyclization, occurs prior to chain extension by Pks5. While SamR0478 has also been speculated to act during chain assembly⁴², hydroxylation evidently occurs downstream of Pks4, at least. The intriguing substrate structural and/or protein–protein recognition features controlling the timing of hydroxylation by these P450 enzymes remain to be elucidated.

Role of TE domains in release of the shunt metabolites

We attributed the observed shunt metabolites to the lack of productive chain translocation between Pks4 and Pks9, causing intermediates to accumulate on ACPs 12 and 13. To evaluate whether these were released by spontaneous hydrolysis or enzymatically, we further investigated the role of the Pks9 TEI⁴² in chain release, as well as that of SamR0485, a proof-reading type II TE⁴³ located in the cluster. Both TEs were disabled by site-directed mutagenesis of the active site serines (Ser to Ala) (Supplementary Fig. 17).

Interestingly, inactivation of both the type I and type II TEs reduced the yields of shunt products 4–7 relative to the parental strain CPN2/OE484 (by 66% and 27%, respectively; average of duplicate experiments) (Supplementary Fig. 18 and Supplementary Table 5). These data clearly show that premature release of the chains is catalyzed, at least in part, by both TEs in the cluster, although spontaneous liberation also occurs. While type II TEs typically interact with acyl-ACPs in trans to release blocked chains⁴³, the effect of the Pks9 TEI is less readily explained. One possibility is that the productive docking interaction between Pks4 and Pks9 allows Pks9 to adopt an alternative conformation from which the TE can off-load intermediates bound to Pks4 ACP₁₂ and ACP₁₃ (Supplementary Fig. 18).

Although this mechanism is reminiscent of that used by the pikromycin PKS to generate both 12- and 14-membered rings⁴⁴, the pikromycin TEI is separated from its alternative ACP target by a single module, while Pks9 TEI is located five or four modules downstream from ACPs 12 and 13 in the engineered system, which would seem to necessitate substantial intersubunit acrobatics. Alternatively, or in addition, such remote off-loading may involve interactions between distinct assembly lines within the context of a PKS megacomplex, as described for the bacillaene system of Bacillus subtilis⁴⁵.

Understanding the docking domain engineering via studies in vitro with recombinant domains

To better understand the results of the DD engineering, we studied in vitro the wild-type DD pairs (^CDD₄/^NDD₅ and ^CDD₈/^NDD₉), as well as binding between the modified versions of ^CDD₄ and wild-type ^NDD₉. Design of suitable expression constructs in E. coli (Supplementary Table 1 and Supplementary Data 1 and 2) was based on bioinformatics analysis of the C-terminal ends of Pks4 and Pks8, and the N-termini of Pks5 and Pks9, and secondary structure analysis using PSIPRED⁴⁶ (Supplementary Fig. 22). Overall, we expressed and purified the following proteins in recombinant form from E. coli: ^CDD₄ wt, ^CDD₄ SDM, ^CDD₄ helix swap, ^NDD₅, and ^CDD₈ (Supplementary Figs. 22 and 23, Supplementary Data 3). As ^NDD₉ proved insoluble when expressed in E. coli, two versions with alternative start sites were obtained as synthetic peptides (Met and Val; Supplementary Fig. 22, Supplementary Table 1). Analysis of the individual ^CDDs by circular dichroism (CD) confirmed their expected high α-helical content (^CDD₄ wt (100 μM): 58%; ^CDD₈ wt (100 μM): 49%), and showed no evident effect of the introduced mutations on secondary structure (Supplementary Fig. 24). All of the constructs were further confirmed to be homodimeric by size exclusion chromatography multi-angle light scattering (SEC-MALS) (Supplementary Fig. 25).

The two ^NDDs also exhibited α-helical character, though less pronounced than the ^CDDs (^NDD₅ (100 μM): 27%; ^NDD₉ Met (100 μM): 21%; ^NDD₉ Val (100 μM): 25%), and were monomeric by SEC-MALS (Supplementary Fig. 25). The latter result was surprising as type 1a ^NDDs classically form a homodimeric coiled-coil domain (Fig. 1, Supplementary Fig. 1), but we recently identified functional, monomeric type 1 ^NDDs⁴⁷. Indeed, we detected binding between the native pairs by isothermal titration calorimetry (ITC), with affinities in the range of those determined previously for matched pairs of DDs^35,47,48,49 (^CDD₄ + ^NDD₅, K_d = 14.5 ± 0.9 μM; ^CDD₈ + ^NDD₉ Met, K_d = 33 ± 2 μM; ^CDD₈ + ^NDD₉ Val, K_d = 22 ± 1 μM) (Supplementary Fig. 26). Thus, while stable homodimerization of the ^NDDs may depend on the presence of a downstream homodimeric KS domain, their monomeric character did not preclude interaction with their ^CDD partners. Based on the higher affinity of the interaction, we could identify the ^NDD₉ Val as the physiologically relevant construct. The observed binding stoichiometry (1 homodimeric ^CDD:2 monomeric ^NDDs), is consistent with the known structure of a type 1a complex in which two monomers of each DD are present (Fig. 1, Supplementary Fig. 1)²⁶. As expected, no nonspecific interaction was detected between native ^CDD₄ and ^NDD₉, explaining the lack of productive communication between subunits Pks4 and Pks9 when the intervening multienzymes are deleted (strain K7N3) (Fig. 3a).

Analysis by ITC of binding between ^CDD₄ SDM or ^CDD₄ helix swap and ^NDD₅ revealed the complete absence of interaction (Supplementary Fig. 26), and therefore that the introduced modifications were sufficient to disrupt communication between the native pair. Thus, the continued production of stambomycins 1 by K7N4, CPN4, and K7N5 harboring Pks5–Pks8 must be due to additional contacts between Pks4 and Pks5 beyond the docking domains, likely including the compatible ACP₁₃/KS₁₄ interface¹⁵. On the other hand, no interaction was detected between ^CDD₄ SDM and ^NDD₉, showing that this limited number of mutations was inadequate to induce productive contacts. This result is fully in accord with the absence of the expected mini-stambomycin products from these strains (K7N1/CPN1, Fig. 3a). By contrast, the ^CDD₄ helix swap exhibited essentially the same binding to ^NDD₉ Val as ^CDD₈ (K_d = 21.0 ± 0.3 μM), demonstrating that exchange of just this helix is sufficient to redirect docking specificity³⁸. Thus, inefficient docking is not at the origin of the failure of the ^CDD₄ helix swaps to yield chain-extended products in vivo (strains K7N2/CPN2, Fig. 3a). We could therefore conclude at this stage that the problem arose from the non-native interface generated between ACP₁₃ and KS₂₁, poor acceptance by KS₂₁ of the incoming substrate during chain transfer and/or chain extension, and/or low activity towards the modified intermediate of domains and modules acting downstream.

Attempted optimization of the stambomycin DD mutants

We aimed next to improve the engineered Pks4/Pks9 intersubunit interface in strain CPN2 (^CDD₄ helix swap + deletion of Pks5–8) by targeting helix αI of ACP₁₃, as the first 10 residues of this helix have been implicated previously in governing the interaction with the downstream KS domain at hybrid junctions⁵⁰. Notably, multiple sequence alignment of all ACPs in the stambomycin PKS located at intersubunit junctions revealed a unique sequence for each ACP in the helix αI region. This observation is consistent with a recognition code for the KS partner, and the idea that mismatching these contacts might hamper productive chain transfer (Supplementary Fig. 27). Indeed, as mentioned previously, even when docking is interrupted, contacts between ACP₁₃ and KS₁₄ are apparently sufficient to enable chain translocation between Pks4 and Pks5 (Fig. 3a). In addition, an analogous strategy of optimizing the ACP_n/KS_n+1 chain transfer interface was shown recently to substantially improve interaction between an ACP (JamC) derived from the jamaicamide B biosynthetic pathway, and the first chain extension module of the lipomycin PKS (LipPKS1)⁵¹.

In our case, the first six residues of ACP₁₃ helix αI were modified using CRISPR-Cas9, so that the full 10-residue recognition sequence matched that of ACP₂₀, the natural partner of KS₂₁ (EADQRR → PSERRQ) (Supplementary Figs. 27 and 28). Analysis of extracts of the resulting strain CPN2/OE484/ACP₁₃ SDM by LC-ESI-HRMS revealed at best small amounts (maximum yield of 0.1 mg L⁻¹) of target cyclic mini-stambomycins A/B (13), lacking the hydroxyl group introduced by SamR0478 (Fig. 4, Supplementary Fig. 29 and Supplementary Table 6). Thus, while this experiment finally yielded evidence for successful chain transfer between Pks4 and Pks9 followed by subsequent chain extension by Pks9 and TE-catalyzed release, the overall efficiency of the system remained low. Interestingly, however, the titers of the four shunt metabolites 4–7 were as much as 48-fold higher from the ACP₁₃ helix swap mutant than from CPN2/OE484. Evidently, improved interactions between ACP₁₃ and KS₂₀ facilitated release of the stalled intermediates from ACPs 12 and 13, presumably via remote action by the TEI domain.

Engineering mini-stambomycins by maintaining the native ACP₁₃/KS₁₄ junction

Cumulatively, the results obtained with the docking domain engineering identified KS₂₁ as one potential bottleneck in the engineered PKS. Our parallel strategy based on PKS XUs (Fig. 2) allowed us to directly test this idea. Specifically, we investigated the effects of preserving the native ^CDD₄/^NDD₅ pair and either the majority of KS₁₄, or a little more than half of the domain, resulting in a KS₁₄/KS₂₁ hybrid. For this, we used two different splice sites in KS₁₄ (Supplementary Fig. 30): (i) at the end of the domain in a highly conserved region (GTNAHV) exploited recently to efficiently swap downstream AT domains⁵²; and, (ii) at a site corresponding to a recombination hot spot identified during induced evolution of the rapamycin (RAPS) PKS⁵³, yielding the KS₁₄/KS₂₁ chimera (Fig. 4 and Supplementary Fig. 31). Both of these modifications were introduced into S. ambofaciens using CRISPR-Cas9, while simultaneously removing Pks5–Pks8, giving, respectively, after co-transformation with pOE484 and the control plasmid pIB139, strains ATCC/OE484/hy59_S1, ATCC/pIB139/hy59_S1, ATCC/OE484/hy59_S2, and ATCC/pIB139/hy59_S2.

Analysis of culture extracts relative to the controls revealed the presence in both ATCC/OE484/hy59_S1 and ATCC/OE484/hy59_S2, of a series of 37-membered metabolites (Fig. 4). The obtained comprehensive MS/MS data were consistent with the desired mini-stambomycins either as their free acids or in cyclic form (metabolites 12–14, Fig. 4, Supplementary Figs. 32–35 and Supplementary Note 1). Signals corresponding to the A/B and C/D derivatives of all metabolites were detected, providing important evidence for their identities, as well as both the C-14 hydroxylated 14 and non-hydroxylated 13 forms of the cyclic mini-stambomycins (C-14 corresponds to C-28 in the parental compounds (Fig. 1)). For detailed justification of the structure assignments of 13 and 14, see the Supplementary Note 1. It is not surprising that the corresponding E and F forms were not detected, as their yields even from the wild-type are much lower than the A–D derivatives (Fig. 3a). Critically, we obtained additional support for the identities of 12–14 by inactivation of samR0479 (which introduces the hydroxyl used for macrocyclization), which resulted in exclusive production of linear deoxy mini-stambomycins 15 (Supplementary Figs. 36–38 and Supplementary Table 7). The observation of non-hydroxylated 13 shows notably that internal hydroxylation by SamR0478 is not an absolute prerequisite for TE-catalysed macrolactonization, and argues that hydroxylation of the mini-stambomycins only takes place on the macrocyclic compound. Although compounds 13 and 14 likely incorporate the tetrahydropyran moiety of the parental stambomycins 1, which undergoes glycosylation, derivatives bearing β-D-mycaminose were not observed, presumably due to poor recognition of the overall modified macrocycle by glycosyl transferase SamR0481³³.

The combined, estimated maximum yields of the target compounds were reduced relative to the wild-type stambomycins by some 8-fold, and variable between fermentations. Notably, however, metabolites 12–14 were obtained at approximately three-fold higher titer from ATCC/OE484/hy59_S2 incorporating the hybrid KS₁₄/KS₂₁ (0.76, 2.0, and 0.74 mg L⁻¹, respectively, (3.5 mg L⁻¹ total)) than from ATCC/OE484/hy59_S1 containing the full KS₁₄ swap (0.27, 0.66, and 0.20 mg L⁻¹, respectively (1.1 mg L⁻¹ total)) (Fig. 4, Supplementary Fig. 32 and Supplementary Table 6). As observed previously, the strains also produced substantial quantities of the shunt products 4–7, while inactivation of samR0479 led correspondingly to the deoxy versions of these compounds 8–11 (Supplementary Figs. 32 and 36). The yields of the shunts were ca. 80-fold higher than those of the corresponding mini-stambomycins, with the highest titers observed in the strain incorporating the hybrid KS₁₄/KS₂₁. The amount of shunt metabolites was also ~123-fold higher than from strain CPN2/OE484 (which incorporates an ACP₁₃–^CDD₄ swap/^NDD₉-KS₂₁ interface) (Figs. 3b and 4, Supplementary Table 6). Thus, contrary to expectation, although using the KS as a fusion site improved communication between Pks4 and Pks9, it also substantially boosted TEI-mediated off-loading of stalled upstream intermediates.

In principle, such stalling could result from a slow rate of chain extension in the now hybrid acceptor module (for example, in the full KS swap construct, KS₁₄ and ACP₂₁ are completely mismatched for chain extension). To evaluate this idea, we modified ACP₂₁ within ATCC/OE484/hy59_S1 incorporating the full-length KS₁₄, targeting a sequence region previously identified as mediating intramodular communication between the KS and ACP during chain extension (Supplementary Fig. 39)^23,50. Specifically, we exchanged loop 1 and the initial portion of helix αII of ACP₂₁ for the corresponding sequence of ACP₁₄, using CRISPR-Cas9 (Supplementary Fig. 39). As we anticipated that creation of this substantially hybrid ACP might engender structural perturbation, we also engineered a minimal mutant of ACP₂₁ in which only one of the two most critical residues in the recognition motif was mutated to the corresponding amino acid in ACP₁₄ (G₁₄₉₉ of Pks9 → D; the second residue, R, of the motif is already common to the two ACPs) (Supplementary Fig. 39). Analysis of the loop/helix αII swap by HPLC-MS showed that all mini-stambomycin production had been abolished (Supplementary Fig. 40 and Supplementary Table 6), consistent with the anticipated disruption to ACP₁₄ structure. Production by the ACP site-directed mutant was not any better than by the full KS swap construct (Fig. 4, Supplementary Fig. 40 and Supplementary Table 6), as only metabolite 13 remained detectable.

In principle, the hybrid KS₁₄/KS₂₁ domain may have worked better than KS₁₄ for chain extension due to improved interaction with ACP₂₁, with stalling displaced to later modules. If this were the case, we might expect to see accumulation in the medium of shunt metabolites corresponding to the intermediate generated by module 21. Indeed, in the case of strain hy59_S2 (chimeric KS₁₄/KS₂₁) but not hy59_S1 (KS₁₄), we detected masses consistent with the A/B and C/D forms of intermediate 16 generated by module 21, at yields comparable to those of the final mini-stambomycins (Fig. 4, Supplementary Fig. 41 and Supplementary Table 6)). Correspondingly, 17, the C30-deoxy analogue of 16, was detected in the SamR0479 mutant (Supplementary Figs. 37 and 38 and Supplementary Table 7). The same metabolite 16 was identified from the ACP₂₁ G →D mutant (Fig. 4 and Supplementary Fig. 40 and Supplementary Table 6), consistent with interrupted chain transfer to KS₂₂. Taken together, these data confirm module 22 as a blockage point in the engineered systems.

Relative efficacy of PKS engineering using PCR-targeting and CRISPR-Cas9

As multiple of our core constructs were generated by both PCR-targeting and CRISPR-Cas9, we were able to directly compare the efficiency of the two techniques (Fig. 3 and Supplementary Figs. 5 and 6). Globally, our results confirm that both approaches can be employed to introduce large-scale modifications to PKS biosynthetic genes (i.e., deletions of single or multi-gene regions)^40,54,55,56. We have also demonstrated that CRISPR-Cas9 can be leveraged to specifically modify modular PKS domains⁵⁷. Of the two methods, CRISPR-Cas9 was the more rapid, as the corresponding constructs were engineered in approximately half of the time. In addition, while CRISPR-Cas9 allowed for direct modification of the host genome, PCR-targeting relied on the availability of suitable cosmids housing the target genes, and resulted in a 33 bp attB-like scar sequence in the genome (Supplementary Fig. 4)⁵⁸. In addition to hampering iterative use of the approach, the scar apparently provoked a moderate reduction in stambomycin yields in mutant K7N6 compared to the wild-type, an effect also noted upon comparison of several analogous mutant strains (e.g., K7N4 vs. CPN4, Fig. 3). Nonetheless, we did encounter certain difficulties with use of CRISPR-Cas9 (i.e., failure to obtain construct CPN3, occasional reversions to wild-type, etc.), observations motivating ongoing efforts in other laboratories to further enhance the suitability of CRISPR-Cas9 for editing PKS pathways^{57,59,60,61,62,63,64}.