Increased Proteolytic Activity of Serratia marcescens Clinical Isolate HU1848 Is Associated with Higher eepR Expression

To better characterize the proteolytic activity displayed by S. marcescens strains, we examined the electrophoretic patterns of their secreted proteins in-gel zymography (Fig. 1). Culture supernatants of strains HU1848, SmUNAM836, and Db10 grown at 37°C were filtered and concentrated through ammonium sulfate precipitation (as described in Material and Methods). Previous analysis of concentrated S. marcescens supernatants from saturated cultures reported null or rare bacterial lysis (Di Venanzio et al. 2014; Lazzaro et al. 2017). An equivalent amount of protein was separated by SDS-PAGE and stained with Coomassie brilliant blue. In agreement with previous reports (Brothers et al. 2015), protein bands above the 48-kDa-protein marker were observed in the supernatant samples of the three S. marcescens strains (Fig. 1A). Clear zones due to hydrolysis of the substrate (gelatin) appeared in the zymogram at the equivalent migration area (above 48 kDa) (Fig. 1B). Moreover, a slightly higher (~55 kDa) second gelatin degradation zone was observed in supernatant samples of HU1848 and Db10 strains, but not in SmUNAM836 (Fig. 1B). Serratia zymography pattern is poorly defined, nonetheless obtained results with HU1848 and Db10 resemble zymogram data of pioneer work in S. marcescens strain BG (Lyerly and Kreger 1979). In contrast, a unique proteolytic zone (similar to SmUNAM836) was also described in the supernatant of the environmental strain AD-W2 (Chander et al. 2021).

Fig. 1.

In order to correlate the protease identity with observed substrate degradation zones, SDS-PAGE-separated bands of HU1848 and SmUNAM836 were excised and subjected to LC-MS/MS. As expected, the protein migrating just above the 48 kDa marker was identified as PrtS (52.2 kDa) in both HU1848 and SmUNAM836, with a peptide coverage of 51% and 38%, respectively (Fig. S1). In addition, SlpB (50.3 kDa) was also identified in the same sample (~48 kDa) of strain HU1848, albeit with low coverage (12%) (Fig. S1). This was in agreement with a report of strain K904, a keratitis isolate, indicating an equal in-gel migration of PrtS and SlpB, as well as lower protein levels of SlpB (Shanks et al. 2015). On the other hand, the SmUNAM836 genome does not encode SlpB.

Moreover, in the 55 kDa hydrolytic zone of HU1848, PrtS was also identified by mass spectrometry with a coverage of 43%. Based on prtS gene sequence, a precursor form of PrtS with an extended N-terminus is considered to occur (Braunagel and Benedik 1990). In addition, the same situation of two closely migrating proteases identified as PrtS after amino-terminal sequencing was previously described in S. marcescens ATCC^® 25419™ (Schmitz and Braun 1985). Synthesis of PrtS as a higher mass pre-protein is also supported by two independent works showing that expression of prtS in E. coli cells results in a slightly larger molecular weight enzyme than the PrtS detected in Serratia (Nakahama et al. 1986; Braunagel and Benedik 1990). No detection of PrtS isoform in SmUNAM836 by zymography might result from nucleotide differences upstream to the putative prtS start codon (data not shown).

To continue characterizing the protease production of S. marcescens isolates, their proteolytic activity was assessed through azocasein hydrolysis using supernatant samples from cultures at 37°C. Obtained results showed a significantly higher (5.4 fold) proteolytic activity of HU1848 compared to SmUNAM836 (Fig. 1C). Protease activity of HU1848 was comparable to the entomopathogen strain Db10 (Fig. 1C). Then, to correlate protease activity determination with protease gene expression, we extracted RNA from the three strains grown at 37°C and achieved qRT-PCR evaluations (Fig. 2). In agreement, prtS transcript levels (Fig. 2A) were significantly elevated in HU1848 (7.24-fold) and in Db10 (5.85-fold) compared to SmUNAM836 (p = 0.028). In contrast, no significant differences were obtained in transcript levels of the protease slpD (Fig. 2B). PrtS is defined as the primary contributor to S. marcescens proteolytic activity (Bruna et al. 2018). Accordingly, the deletion of slpB only showed a slight reduction in the proteolytic capability of the keratitis isolate K904 (Shanks et al. 2015). Thus, we associated the reduced azocasein degradation in SmUNAM836 with lower production of PrtS rather than with the lack of SlpB.

Fig. 2.

To date, prtS expression has been reported to be directly influenced by transcriptional regulators cpxR, eepR, and hexS. (Shanks et al. 2017; Bruna et al. 2018). Therefore, we decided to compare the transcript levels of these three regulators. Our data indicated a 16.8 and 15.9-fold eepR expression in HU1848 and Db10 strains, respectively, related to SmUNAM836 (p = 0.019 and p = 0.015, respectively) (Fig. 2C). In contrast, no significant differences were determined in transcript levels of cpxR or hexS (Fig. 2D and 2E). In order to corroborate the higher eepR transcription determined in HU1848, we fused the eepR upstream region of HU1848 and SmUNAM836 (Fig. 3A) to the promoterless luxCDABE operon in plasmid pSEVA246 (Martínez-García et al. 2020). The S. marcescens strains carrying eepR transcriptional reporter were grown at 37°C and relative luminescence was determined at different time points (Fig. 3B). In agreement with our qRT-PCR data, luminescence values showed a significantly elevated eepR expression in HU1848 strain throughout culture growth compared to SmUNAM836 (p < 0.05) (Fig. 3B). Control plasmids indicated negligible luciferase basal activity in both strains (Fig. 3B), confirming the higher transcriptional activity of eepR in strain HU1848.

Fig. 3.

Our data suggest that the higher proteolytic activity, displayed by HU1848 and Db10 strains, is at least partially associated with elevated eepR expression. Thus, we looked at the eepR upstream sequences. Several substitutions were noticed, including two at the CRP binding site and six at the immediate upstream region (Fig. S2). Comparison with different S. marcescens isolates revealed that these nucleotide changes are shared by several representative strains (Fig. S2). Moreover, two conserved putative CpxR binding sites were identified within eepR regulatory region (Fig. 3A and S2).

CRP has been described as an eepR repressor (Stella et al. 2015; Shanks et al. 2017); based on the position of the sequence differences within the eepR regulatory region of the studied strains, we cloned crp and purified a recombinant CRP (Fig. S3A). However, when we evaluated the interaction of CRP with the eepR upstream sequence of SmUNAM836 or HU1848 a similar migration was observed (Fig. 3C). A comparable result was obtained using a recombinant HexS (Fig. S4). Interaction specificity was corroborated using a DNA fragment lacking the predicted CRP binding site or an unrelated DNA fragment (Fig. S4). According to these results, the differences in eepR expression of studied strains are unlikely as a consequence of CRP or HexS impaired recognition of eepR regulatory region. Nonetheless, a more detailed genetic analysis is needed to assess if the eepR upstream sequence differences (Fig. S2) modify the affinity of CRP or HexS (or another transcriptional regulator), resulting in differential expression of eepR in S. marcescens strains.

In addition, other factors not necessarily defined by the eepR regulatory sequence could indirectly impact eepR transcriptional activity; for instance, unlike SmUNAM836, strains Db11 and HU1848 lack the luxI gene responsible for N-acyl-L-homoserine lactone (AHL) molecule, implicated in quorum sensing (QS) signaling (Sakuraoka et al. 2019). In this regard, QS signaling in the environmental species Serratia liquefaciens MG1 and Serratia proteamaculans B5a was found to be involved in the regulation of the T1SS lipBCD and consequently affecting the exoenzymes secreted through this system (Riedel et al. 2001; Christensen et al. 2003), a phenotype that resembles a S. marcescens eepR deletion strain (Brothers et al. 2015; Stella et al. 2017). Nonetheless, a direct impact of QS molecules over eepR expression in AHL producer Serratia strains remains to be investigated.

Two putative CpxR binding sites (greatly conserved between S. marcescens species) were noticed within eepR upstream sequence (Fig. 3A and S2). To assess whether CpxR directly contributes to the expression of eepR we purified a recombinant CpxR protein (Fig. S3C). It was implemented in EMSA evaluations using different DNA fragments containing or not the putative CpxR binding sequences (Fig. 3A and 3D). A clear interaction of CpxR to eepR upstream sequence was noticed, particularly when both binding sites were included (Fig. 3D; E1 lane). CpxR interaction decreased when a single site was present (Fig. 3D; E2 and E3 lanes), indicating that CpxR binds to both motifs. Lastly, CpxR interaction appears specific since it was lost when a fragment lacking both recognition sequences was tested (Fig. 3D; E4 line). In addition, CpxR binding was similar in the absence of acetyl phosphate, suggesting that CpxR can interact in vitro with eepR regulatory region regardless of its phosphorylation (Fig. 3D).

Moreover, to evaluate the role of CpxR over transcriptional regulation of eepR, a cpxR deletion strain was constructed in the HU1848 background. Total RNA was extracted from cultures at 37°C, a temperature at which CpxR became more active (Bruna et al. 2018). We found that eepR expression was 2.3-fold elevated in the cpxR mutant (Fig. 3E, p = 0.03) compared to the parental strain. Also, the eepR mRNA values were lowered when the cpxR mutant was carrying the cpxR gene in trans (Fig. 3E). Therefore, our data confirms a direct role of CpxR operating as a negative regulator of eepR.

According to its direct positive role in producing different exoenzymes, including PrtS (Stella et al. 2015), EepR is considered a key regulator of host-pathogen interactions promoting proinflammatory response (Brothers et al. 2021). In this scenario, the induction of prtS might further contribute to bacterial airway pathogenicity through the activation of the epithelial sodium channel (ENaC) (Butterworth et al. 2014). Also, experimental models have shown that intranasal administration of PrtS provokes a deleterious impact, leading mouse lungs to be markedly susceptible to viral infection (Akaike et al. 1989). Nonetheless, elucidating an early cytotoxicity might accelerate bacterial clearance, preventing tissue colonization. Thus, negative thermoregulation of prtS by CpxR (at 37°C) is believed to grant warm-blooded hosts colonization and to regulate S. marcescens biofilm community (Bruna et al. 2018).

In summary, we have shown here that elevated expression of EepR in the respiratory isolate HU1848 is associated with increased prtS levels and concomitant proteolytic activity. Moreover, we described that CpxR directly binds to the upstream region of eepR and inhibits the expression of this positive regulator. However, further research is needed to understand the molecular mechanisms relieving the eepR repression on particular S. marcescens isolates.