Bacterial strains and plasmids used in this work are listed in Table 1. E. coli JM109 was used for cloning. For conjugation, donor strains were either E. coli BW19851 (λ pir) or E. coli S17 (λ pir) and recipient strains were either E. coli TransforMax EC100+ (for propagation of constructs) or Delftia sp. Cs1-4. E. coli strains were routinely grown in Luria-Bertani (LB) broth at 37°C. Mineral salt medium (MSM; Hickey and Focht, 1990) containing phenanthrene as the sole carbon source (1 mg/mL) was routinely used for Delftia sp. Cs1-4 culture. Liquid cultures were grown with shaking (ca. 200 rpm) at either 25°C (strain Cs1-4) or 37°C (E. coli). For solid LB media, Bacto-Agar (Difco, Detroit, MI, USA) was added to a final concentration of 15 g/L. For E. coli, antibiotics were added when required at 100 μg/mL (ampicillin, Ap), 50 μg/mL (kanamycin, Km), or 10 μg/mL (tetracycline, Tc). Kanamycin and tetracycline were used in some Delftia sp. Cs1-4 cultures, and in these cases were added at 300 and 40 μg/mL, respectively.

Genomic DNA was prepared using a genomic DNA extraction kit (Promega, Madison, WI, USA), and plasmid DNA was purified with the QIAprep spin miniprep kit (QIAGEN, Germantown, MD, USA). Restriction and modification enzymes were purchased from Promega (Madison, WI, USA) or New England Biolabs (Beverly, MA, USA). Klenow fragment or T4 DNA polymerase (Promega) was used to fill in recessed 3′ ends and to trim protruding 3′ ends of incompatible restriction sites. All PCR amplifications were done with the Failsafe PCR system (Epicenter Technology, Madison, WI, USA). Amplicons were separated in 0.7–1.0% (w/v) agarose gels, and DNA fragments were purified with the QIAquick gel extraction system (QIAGEN). Ligation mixtures were transformed into E. coli JM109 (Promega), and transformants were plated onto LB plates with appropriate antibiotic selection. Resistant colonies were isolated, and then screened for the acquisition of plasmids. All constructs were sequenced to verify structure. For conjugal transfer of plasmids from E. coli to Delftia sp. Cs1-4, LB-grown cultures of both cells were harvested (mid-log phase) by centrifugation, washed with LB and then equal amounts (ca. 10¹² cells of each strain) were mixed, and spotted onto LB plates containing 5 mM CaCl₂. Following overnight incubation at 22°C, cells were then scraped off of the plates, diluted, and plated on LB plates containing the appropriate antibiotics.

Total RNA was isolated from phenanthrene-grown strain Cs1-4 cells, and purified of genomic DNA by DNase I digestion. Analysis by 5′-RACE was done using TaKaRa 5′-full RACE Core set under conditions recommended by the supplier (TaKaRa). Reverse transcription (RT) was done with a 5′-phosphorylated RT primer (Delf1; Table 2). After RT, mRNA was digested with RNaseH, and then cDNA was concatenated using T4 RNA ligase. The region of interest was then amplified via nested PCR using two sets of primers to regions of npdA. In the first PCR, RT products were used as template, and amplified with primers Delf2 and Delf3 (Table 2). In the second PCR, template was a 10-fold dilution of the round one PCR product, and amplification was done using primers Delf4 and Delf5 (Table 2). The 5′-RACE products were isolated, purified, ligated into pGEM-T easy and then sequenced.

The npdA fragment including the non-coding and partial structural gene regions was amplified with primers Delf6 and Delf7 (Table 2) using strain Cs1-4 genomic DNA as template. The Renilla luciferase (rluc) gene was amplified from pRL-SV40 using primers Delf8 and Delf9 (Table 2). These fragments were fused via overlap PCR. To analyze the structure of the putative npdA promoter, deletion derivatives of non-coding fragments upstream of npdA were amplified by employing the same PCR strategy as described above, except using different N-terminal primers, namely Delf10, Delf11, Delf12, Delf13, Delf14, and Delf15 (Table 2). The above amplicons were inserted in pGEM-T easy, released from this vector by SacI and SacII digestion, and inserted into the same sites of pBBR1MCS-3 to create the deletion series. The reporter vector was then conjugated into strain Cs1-4.

Genes encoding green fluorescent protein and red fluorescent protein were amplified from pKEN2 and pmStrawberry using the primers Delf16/Delf17 and Delf18/Delf19, respectively, and engineered via PCR to contain an E. coli ribosome binding site on the 5′-end (Table 2). The amplicons were cloned into pGEM-T easy (pSCH374 and pSCH378, respectively), gfpmut3 was then released by ApaI and SacII digestion, and inserted into the same sites on pBBR1MCS3 (pSCH397). The mStrawberry gene was cut from pSC378 by digestion with KpnI and SacII, and inserted into KpnI/SacII sites on pBBR1MCS3 (pSCH395).

A strong expression system controlled by PnpdA was constructed as follows. The PnpdA region (genome position 5862152–5862685) was amplified from strain Cs1-4 genomic DNA using primers Delf20 and Delf21 (Table 2). The amplicon was then cloned into pGEM-T easy (pSCH426), released by digestion with ApaI and SmaI, and inserted into the same sites on pBBR1MCS3 (pSCH442). Green fluorescent protein (GFP, gfpmut3) and red fluorescent protein (RFP, mStrawberry) marker genes were released from pSCH374 and pSCH378 by digestion with SacII, cloned into pSCH442 and transformed into E. coli JM109. Colonies with strong green (pSCH476) and red (pSCH473) fluorescence were recovered, and orientation of reporter genes was confirmed by sequencing. These plasmids were next conjugated into Delftia sp. Cs1-4, leading to strains SCH481 (pSCH476) and SCH482 (pSCH473).

To transcriptionally tag npdA, gfp was inserted immediately downstream of npdA using the Cre-loxP recombination method of Denef et al. (2005). An npdA fragment with the stop codon (genome position 5860670–5861289) was amplified using primers Delf22 and Delf23 (Table 2). The downstream fragment of npdA (genome positions 5860066–5860809) was amplified using primers Delf24 and Delf25 (Table 2). These fragments were then cloned into pGEM-T easy (pSCH447 and pSCH430). The downstream fragment from pSCH430 was released by digestion with SacII and SacI and inserted into the same sites on pJK100 (pSCH431). The npdA fragment from pSCH447 was released by NdeI and KpnI digestion, and then inserted into the same sites on pSCH431 (pSCH485). The gfpmut3 gene was released from pSCH375 by KpnI and NotI digestion, and assembled into the same sites on pSCH485 (pSCH451). Conjugation of pSCH451 into strain Cs1-4 gave Km^r/Tc^s colonies, which were recovered for further analysis. The Cre-expressing vector, pCM157, was next introduced into a selected colony (SCH483) in order to remove K_m resistance, leading to strain SCH484 (Km^s/Tc^r). Curing of pCM157 from SCH484 was done by serial transfers in LB medium. A selected colony (Km^s/Tc^s) with green fluorescence was then confirmed for the correct construct by PCR and sequencing (SCH456).

To knock out lasI, its upstream (strain Cs1-4 genome positions 1950815–1951882) and downstream (strain Cs1-4 genome positions 1952504–1953573) fragments were amplified with primers Delf26/Delf27 and Delf28/Delf29, respectively (Table 2). The amplicons were gel purified and cloned into pGEM-T easy (pSCH488 and pSCH356). The upstream fragments were released by BglII/NdeI digestion, and downstream fragments were released by ApaI/SacI from pGEM-T easy and then sequentially assembled on the same sites on pJK100 (pSCH363). To knock out hcp, upstream (strain Cs1-4 genome position 3366999–3367911) and downstream fragments (strain Cs1-4 genome position 3368229–3369041) were amplified using PCR primers Delf30/Delf31 and Delf32/Delf33, respectively (Table 2). The amplicons were gel purified and cloned into pGEM-T easy (pSCH487 and pSCH486). These fragments were sequentially assembled on the same sites on pJK100 (pSCH339) using the same strategy as described above. To knock out omp32, upstream (Cs1-4 genome positions 1041477–1042202) and downstream (Cs1-4 genome position 1044310–1045032) fragments were amplified with primers Delf34/Delf35 and Delf36/Delf37 (Table 2). The amplicons were gel purified and cloned into pGEM-T easy vector (pSCH490 and pSCH418). These fragments were sequentially assembled on the same sites on pJK100 (pSCH371). Each of the three constructs (pSCH363, pSCH339, pSCH371) was introduced into strain Cs1-4 by conjugation, and Tc^s/Km^r transconjugants were selected, leading to strains SCH369, SCH340, and SCH389, respectively.

Modification of pHimarEm1 was done to introduce additional unique KpnI–BamHI–SacII restriction sites, to remove the erythromycin resistance gene and to insert genes encoding GFP and RFP. To do so, PCR was done with pHimarEm1 DNA as template, and using forward primer Delf38 and reverse primer Delf39 (Table 2). The amplicon was digested with BamHI, self-ligated and transformed into E. coli S17 λpir. The gfpmut3 fragment was digested with KpnI and SacII from pSCH375 and inserted into pSC29 at the same restriction sites (pSCH160). The promoterless mStrawberry fragment was then released from pSCH378 by KpnI and SacII digestion, inserted into pSCH29 at the same restriction sites (pSCH402), and then introduced into strain SCH456 by conjugation. The Km-resistant colonies were randomly picked and replicated in 96-well plates containing MSM with either pyruvate and or phenanthrene as the carbon source. After incubation with shaking (24 h), the OD₆₀₀ and GFP fluorescence were determined (see below).

Renilla luciferase assays were done as described in our prior work (Chen et al., 2009) using a commercially available kit (Promega) according to the manufacturer’s protocol. Quantitative analysis of fluorescent protein production was done using a Synergy 2 plate reader with the following conditions (all 0.2-s interval, 22°C): GFP, excitation at 485 nm, emission 510 nm; RFP, excitation at 574 nm, emission at 596 nm. All measurements were corrected for background with wild type (WT) Delftia sp. Cs1-4 cells.

The complete genome sequence of Delftia sp. Cs1-4 was deposited in Genbank as accession NC(015563.1. All constructs were sequenced by the dideoxy termination method using an Applied Biosystems (Foster City, CA, USA) 3730 × l DNA Analyzer available at the University of Wisconsin-Madison, Biotechnology Center. GenBank database searches were carried out using the National Center for Biotechnology Information BLAST-N web server.

Three TSS were identified for npdA, and were located at (nucleotide) −34-bp (A), −56-bp(G), and −172-bp (A), respectively upstream of the npdA start codon (Figure 1A). Three putative promoter motifs, PnpdA₁ (TCCTCT-N₁₅-TGTCTG), PnpdA₂ (TAGGGG-N₁₅-TACGAT), and PnpdA₃ (TACGAT-N₁₇-TGGTGG) situated at −38, −61, and −180-bp, respectively were identified (Figure 1A). Serial deletion of non-coding regions upstream of npdA was done to establish involvement in npdA regulation of one or more of the three putative promoters. There was no significant difference in levels of gene expression between the WT and D1 (npdA −220 bp; Figure 1B). However, further deletion of an 11-bp fragment from D1 (D2, npdA −209 bp) yielded a ca. 20% decrease in Rluc activity relative to the WT (Figure 1B). Since the D2 construct carried the putative −35 motif in PnpdA₁, we inferred the fragment (−220 to −209 bp) was also important for npdA expression. Deletion of the −35 region of PnpdA₁ (D3, npdA −190 bp) decreased Rluc activity by >40% compared to the WT. Construct D4 (npdA −180 bp) had only ca. 20% Rluc activity. The latter contained a deletion that originated at −180 bp, and thus had the entire PnpdA₁ region disrupted, indicating that PnpdA₁ was the most important promoter for driving npdA expression. A further deletion (D5, npdA −67 bp) that removed the –35 bp motif in PnpdA₂ retained ca. 5% of WT level. Removing the PnpdA₂ region (D6, npdA −54) reduced Rluc activity to background levels.

To test the utility of the PnpdA expression system, the genes encoding a GFP and RFP were inserted downstream of the PnpdA cassette, which contained the 220-bp fragment described above. Transformants appeared green or red under ambient light, indicating strong expression of gfp and mstrawberry, respectively. The apparent high-level expression of these proteins was non-toxic to Delftia sp. Cs1-4, as growth of cultures expressing GFP or RFP was not distinguishable from that of the WT (Figure 2A). Production of GFP and RFP followed similar patterns, with levels increasing with culture growth, achieving stable accumulations upon reaching stationary phase (Figure 2B). In the absence of antibiotic selection, the expression vector was stable in Delftia sp. Cs1-4 for at least 56 generations (Figure 2C).

For generation of gene knockouts, the vector was used to target omp32, hcp, and lasI. Deletion of all three genes was successful, and confirmed by PCR and/or Southern hybridization. However, none of the gene deletions resulted in a loss of nanopod production, and only the Δomp32 mutant exhibited phenotypes different from that of the WT. In whole cell protein profiles, the latter mutant showed a loss of the predominant band corresponding to Omp32 (Shetty et al., 2011) and appearance of two other proteins, also identified as porins (Figure 3A). The Δomp32 mutant had an irregular cell shape (Figure 3B), and its growth was impaired on both pyruvate and phenanthrene, but the impact of Omp32 loss appeared to be greater with the latter substrate (Figures 3C,D).

Following conjugal delivery to Delftia sp. Cs1-4, the transposition frequency of pMiniHimar was ca. 2 × 10⁻⁵ to 5 × 10⁻⁶ per recipient, a frequency comparable to those reported for Shewanella oneidensis, Geobacter sulfurreducens, and B. pseudomallei (Choi et al., 2008; Rollefson et al., 2009). From the 13,000 colonies screened, seven mutants were recovered that were impaired in either growth on phenanthrene (Mutants 1–6; Table 3) or in npdA expression (Mutant 7; Table 3). For the former, three mutants had insertions in the gene cluster encoding the phenanthrene catabolic pathway. Of these, Mutant 3 was intriguing as the gene bearing the insertion was predicted to encode an Ycf48 homolog. For Mutants 5 and 7, insertions were in genes outside of the phenanthene degradation cluster, and were predicted to encode a SpoT/RelA-type (p)ppGpp synthetase, and a HylD Family, type I secretion membrane fusion protein, respectively.