Complementation of Cold Shock Proteins by Translation Initiation Factor IF1 In Vivo

Abstract

The cold shock response in both Escherichia coli and Bacillus subtilis is induced by an abrupt downshift in growth temperature and leads to a dramatic increase in the production of a homologous class of small, often highly acidic cold shock proteins. This protein family is the prototype of the cold shock domain (CSD) that is conserved from bacteria to humans. For B. subtilis it has been shown that at least one of the three resident cold shock proteins (CspB to D) is essential under optimal growth conditions as well as during cold shock. Analysis of the B. subtilis cspB cspC double deletion mutant revealed that removal of these csp genes results in pleiotropic alteration of protein synthesis, cell lysis during the entry of stationary growth phase, and the inability to differentiate into endospores. We show here that heterologous expression of the translation initiation factor IF1 from E. coli in a B. subtilis cspB cspC double deletion strain is able to cure both the growth and the sporulation defects observed for this mutant, suggesting that IF1 and cold shock proteins have at least in part overlapping cellular function(s). Two of the possible explanation models are discussed.

Most if not all organisms studied to date respond to a sudden decrease in temperature by induction of a complex cascade of adaptation reactions that, in summary, has been termed the cold shock response (4, 8, 11, 16, 48). In Escherichia coli and Bacillus subtilis, this process is characterized by specific alteration of protein synthesis in order to cope with cold-associated problems affecting the proper function of membrane (26, 46), metabolism (12, 23) and, most importantly, the translation apparatus (21, 22). It has been shown that the most dramatically cold-induced proteins (CIPs) belong to an ancient family of small (ca. 70 amino acid residues), often highly acidic proteins that seems, with only a few exceptions, universally conserved among bacteria (10, 13). These proteins are designated cold shock proteins (CSPs) and have also been identified as subdomains in a variety of important eucaryotic proteins involved in coupling of transcription to translation (45). Determination of the three-dimensional structures of CspB from B. subtilis (37, 40), CspA from E. coli (34, 38), and CspB from the thermophilic Bacillus caldolyticus (33), as well as numerous biochemical studies, revealed that CSPs have a conserved unique structural fold that binds single-stranded nucleic acids with variable binding affinities and sequence selectivities depending of the individual protein examined (9, 27, 28, 35, 43). Some but not all of these proteins have been shown to form dimers in vitro (29, 31, 37, 51). Therefore, in spite of a very similar overall fold, these individual differences in CSP:CSP and CSP:RNA/DNA interactions identified in vitro may also reflect different functions in vivo, a suggestion that would explain why csp genes are found in differentially regulated multiple copies of varying homologies within a given bacterial species (49). Indeed, it has been demonstrated that distinct CSPs are involved in different functions, with some of them having more than only one. Apart from its properties as a transcriptional activator (2), CspA from E. coli has also been shown to act as an RNA chaperone (20) and possesses transcriptional antiterminator functions like its homologs CspC and CspE (1). Moreover, CspE was found to be implicated in promoting or protecting chromosome folding and to act as a high-copy suppressor of mutations in the chromosomal partition gene mukB (19), while CspD, which appears to exist exclusively as a homodimer, is specifically expressed in the stationary phase and has been shown to inhibit replication (50, 51).

However, although CSPs were originally identified as the major CIPs, for B. subtilis it has become evident that this essential protein class is generally required for survival even under optimal growth conditions (13). This indicates a fundamental, yet unknown cellular function of B. subtilis CSPs. Interestingly, it is the initiation of translation that is most affected by a drastic reduction in growth temperature (3, 6, 7, 21). This process is believed to represent the bottleneck of cold shock adaption in microorganisms. Besides other factors, in E. coli, the essential translation initiation factor IF1 has been demonstrated to be an important component involved in this crucial step of protein biosynthesis (reference 14 and references therein). In contrast to E. coli, for B. subtilis production of an IF1 protein homolog has not been reported so far. Therefore, on the basis of the recently determined three-dimensional structures of IF1 from E. coli (44) and of CspB from B. subtilis (37, 40), which show an almost identical fold (Fig. 1), we wondered whether IF1 would be able to functionally complement CSPs in vivo. In this work, we show that heterologous expression of IF1 from E. coli in B. subtilis can indeed cure both the growth and the differentiation defects observed in a B. subtilis cspB cspC double-deletion mutant.

FIG. 1.

Comparison of the structures of cold shock protein CspB from B. subtilis (A, PDB accession no. 1CSP [37]) and IF1 from E. coli (B, structure 4 of PDB accession no. 1AH9 [44]). Ribbon models were generated with Swiss-PdbViewer version 3.7b2 (15), and images were rendered by using POV-Ray version 3.1g. (C) Superimposition of IF1 (blue) on CspB (yellow) was performed by using the “iterative magic fit” function of Swiss-PdbViewer.

Bacterial strains, media, and growth conditions.

All strains used in this study are described in Table 1. E. coli GM48 and E. coli XL1-Blue were used as cloning hosts for plasmid construction and were grown in Luria-Bertani (LB) medium (36) supplemented with 50 μg ampicillin per ml where appropriate.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Genotypea	Source or reference
Strains
B. subtilis 64BC	JH642 ΔcspB::cat::spc(Spr) ΔcspC::kan(Kmr)	13
B. subtilis 64BCDbt	JH642 ΔcspB::cat::spc(Spr) ΔcspC::kan(Kmr) ΔcspD::cat(Cmr) pDGcspB	13
B. subtilis 64BD	JH642 ΔcspB::cat::spc(Spr) ΔcspD::cat(Cmr)	13
B. subtilis 64CD	JH642 ΔcspC::kan(Kmr) ΔcspD::cat(Cmr)	13
B. subtilis JH642	pheA1 trpC2 sfp0	18
B. subtilis MW_pDG148	JH642 pDG148(Nmr Pmr)	This work
B. subtilis MW_IF1Ec	JH642 pMW_infAEc-1(Nmr Pmr)	This work
B. subtilis MW_ΔCD-IF1Ec	64CD pMW_infAEc-1(Nmr Pmr)	This work
B. subtilis MW_ΔBC-IF1Ec	64BC pMW_infAEc-1(Nmr Pmr)	This work
E. coli XL1-Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F′ proAB lacIqZΔM15 Tn10(Tetr)]	Stratagene
E. coli GM48	ara dam dcm galK galT leu supE44 thi-1 thr tonA tsx	52
Plasmids
pDG148	PspaclacI bla neo(Nmr) phleo(Pmr)	Patrick Stragier
pMW_infAEc-1	pDG148 infA	This work

^aSp^r, spectinomycin resistance; Km^r, kanamycin resistance; Cm^r, chloramphenicol resistance; Nm^r, neomycin resistance; Pm^r, phleomycin resistance; Tet^r, tetracycline resistance. 

B. subtilis strains JH642, 64BC, 64BCDbt, 64BD, 64CD, MW_pDG148, MW_IF1^Ec, MW_ΔCD-IF1^Ec, and MW_ΔBC-IF1^Ec were grown in LB medium for preparation of chromosomal DNA and PCR analysis. The antibiotics chloramphenicol, kanamycin, spectinomycin, or phleomycin were added as supplements where appropriate.

All growth measurements were carried out at 37°C in SMM^TrpPhe supplemented with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) as described previously (46) and are the results of at least three independent experiments. Spore assays were performed twice in DSM sporulation medium containing 1 mM IPTG according to standard protocols by using the heat-kill method (17).

Construction of plasmids and strains.

The infA gene from E. coli XL1-Blue, encoding translation initiation factor IF1, was amplified by PCR (Table 2) introducing a ribosomal binding site optimized for B. subtilis using PCRMix3_infA^Ec and was cloned into shuttle vector pDG148 by using restriction sites HindIII and XbaI to give plasmid pMW_infA^Ec-1. After passage through E. coli XL1-Blue, this plasmid was verified by sequencing and was transformed into B. subtilis 64CD essentially according to the method reported by Klein et al. (25). The resulting strain B. subtilis MW_ΔCD-IF1^Ec was transformed with chromosomal DNA from B. subtilis 64BC and selected on LB plates containing 5 μg of phleomycin and 100 μg of spectinomycin per ml. Transformants were successfully tested by PCR specific for the presence of the pDG148 derivative pMW_infA^Ec-1 (PCRMix3_pDG148) and the three B. subtilis csp genes cspB (PCRMix3_ΔcspB), cspC (PCRMix3_ΔcspC), and cspD (PCRMix3_ΔcspD; for details, see Table 2 and the legend of Fig. 3). Among the pMW_infA^Ec-1-carrying colonies tested, one carried deletions of cspB and cspC, while cspD had been restored during the transformation process (see Fig. 3). This strain was designated B. subtilis MW_ΔBC-IF1^Ec, and was used for further studies.

TABLE 2.

PCR mixtures, primers, annealing temperatures, and elongation periods used in this study^a

PCR mixture (T1, T2, t)	Primer	Sequence (introduced cutting site)
PCRMix3_infAEc (55, 65, 0:35)	5′infAec-SDBs	5′-GATATAAGCTTAGGAGGAAATATCATGGCGAAAGAAGACAATATTGAAATGCAAGG (HindIII)
	3′infAec-STOP	5′-GACATTCTAGATCAGCGACTACGGAAGAC (XbaI)
PCRMix3_ΔcspB (48.5, 62, 2:40)	5′cspB-TLProm	5′-GGAAGAATTCTTCCTATTACTTATCTTCC (EcoRI)
	3′cspB-TLTrm2	5′-AAAAGGATCCGATCTACTTTATCATCTTC (BamHI)
PCRMix3_ΔcspC (48.5, 62, 2:40)	5′cspC-TLProm	5′-GGAAGAATTCGATTAATTTGCTGGAAATG (EcoRI)
	3′cspC-TLTrm	5′-AAAAGGATCCGTAGGGTTTTTTATTGAGT (BamHI)
PCRMix3_ΔcspD (48.5, 62, 2:40)	5′cspD-TLProm	5′-GGAAGAATTCTGAGAGCATACGAAAATC (EcoRI)
	3′cspD-TLTrm	5′-AAAAGGATCCTGAAATCCAGATGAAAATC (BamHI)
PCRMix3_pDG148 (47.5, 61, 1:56)	5′pDG148-phleo	5′-AAAAGCCGGCGACAAAACCACTCAAAATA (NaeI)
	3′pDG148-oriBs	5′-AAAAGCCGGCTTGATACATATAGAAATAACGT (NaeI)
PCRMix3_Pspac-infAEc (53, 53, 0:40)	5′Pspac	5′-GGCATAATGTGTGGAATTGTG
	3′infAec-STOP	5′-GACATTCTAGATCAGCGACTACGGAAGAC (XbaI)
PCRMix3_Pspac-cspB (53, 53, 0:40)	5′Pspac	5′-GGCATAATGTGTGGAATTGTG
	3′cspB-gfp	5′-AAACCCGGGACGCTTCTTTAGTAACGT (SmaI)

^aT₁, primary annealing temperature (°C); T₂, secondary annealing temperature (°C); t, elongation period (min:s). Restriction sites are underlined 

FIG. 3.

PCR verification of the constructed B. subtilis strain MW_ΔBC-IF1^Ec. The specific DNA regions analyzed are indicated within the figure (cspB, cspC, cspD, pDG148, P_spac-cspB, and P_spac-infA), and lanes are identified as follows: 1, PCR analysis of parental strain JH642 harboring all three csp genes (cspB to -D); 2, PCR analysis of control strain 64BCDbt harboring no chromosomal csp gene but a copy of cspB on the pDG148 derivative plasmid pDGcspB in trans; 3, PCR analysis of the cspB cspC double deletion mutant MW_ΔBC-IF1^Ec harboring the E. coli infA gene on the pDG148 derivative plasmid pMW_infA^Ec-1; M, HindIII-digested λ-DNA was used as a marker. For PCR analysis of wild-type cspB, cspC, and cspD gene loci, amplificate sizes of ca. 0.5 kb each were expected (see JH642 wild-type control), whereas a deletion in cspB, cspC, and cspD should yield fragments of ca. 3, 1.5, and 2 kb, respectively (see 64BCDbt deletion control). For the presence of a pDG148 derivative plasmid, a PCR amplificate of ca. 3 kb was expected, and for the presence of the cspB or infA gene located downstream of a P_spac promotor, PCR amplificates of approximately 0.3 kb were expected. PCR conditions are detailed in Table 2.

To construct the control strains B. subtilis MW_pDG148 and MW_IF1^Ec, parental strain B. subtilis JH642 was transformed with purified pDG148 and pMW_infA^Ec-1, respectively. Transformants were selected on LB plates supplemented with 5 μg of phleomycin and tested for presence of pDG148 derivatives by PCR as stated above (data not shown).

All PCR analyses were carried out in a Perkin-Elmer GeneAmp PCR System 9700 by using the Expand Long Range PCR Kit from Boehringer Mannheim essentially according to the protocols supplied by the manufacturer. Sequencing was performed in an ABI Prism 310 Genetic Analyzer from PE Applied Biosystems by using the ABI Prism dRhodamine Terminator Cycle Sequencing Ready Reaction Kit as suggested by PE Applied Biosystems.

Complementation of CSPs by translation initiation factor IF1.

To test the ability of translation initiation factor IF1 from E. coli to fulfill functions of the B. subtilis cold shock proteins in vivo, a B. subtilis strain was constructed as described that carried deletions in the cspB and cspC genes and harbored a multicopy plasmid allowing expression of E. coli infA encoding for translation initiation factor IF1. The resulting strain MW_ΔBC-IF1^Ec was subjected to growth studies in comparison to B. subtilis 64BC, which differs from MW_ΔBC-IF1^Ec only in lacking the IF1 expressing plasmid. As shown in Fig. 2, MW_ΔBC-IF1^Ec shows a higher doubling rate than 64BC (13) during exponential growth if grown in glucose minimal medium at 37°C. In addition, strain MW_ΔBC-IF1^Ec in contrast to 64BC does not undergo lysis after entry into stationary growth phase. Control growth experiments carried out with B. subtilis strains JH642 (parental strain), MW_pDG148 (JH642 carrying parental vector pDG148), and MW_IF1^Ec (JH642 carrying the infA harboring pDG148 derivative pMW_infA^Ec-1) revealed no detectable difference compared to MW_ΔBC-IF1^Ec (Fig. 2). These results demonstrate that, under the conditions tested, (i) an infA expressing cspB cspC double mutant behaves like wild-type B. subtilis and (ii) infA expression has no detectable effect on the growth of the wild type.

FIG. 2.

Growth comparison of B. subtilis strains JH642 (Δ, parental strain, no phleomycin added), MW_ΔBC-IF1^Ec(■, IF1 expressing cspB/cspC double-deletion strain, 5 μg of phleomycin per ml added), and 64BC (□, IF1 nonexpressing cspB/cspC double-deletion strain, no phleomycin added) grown in glucose minimal medium supplemented with 50 μg of tryptophane per ml, 50 μg of phenylalanine per ml, and 1 mM IPTG. Mean values and standard deviation of three independent experiments are shown.

To further examine the ability of the IF1 expressing cspB cspC double-deletion strain to differentiate into endospores, spore assays were carried out as described. In contrast to strain 64BC which fails to perform this complex differentiation process due to a block at sporulation stage 0 (47), MW_ΔBC-IF1^Ec is able to sporulate with efficiencies comparable to those of strain JH642. While 55% of JH642 cells viable after 24 h of incubation in DSM at 37°C differentiated into endospores, under the same conditions 54% of MW_ΔBC-IF1^Ec cells had sporulated. Spores of strain MW_ΔBC-IF1^Ec outgrown on LB plates supplemented with 5 μg of phleomycin per ml were again verified by PCR analysis as stated above and were successfully shown to harbor the two csp gene deletions and the infA expression plasmid as expected (data not shown).

To test whether the presence of IF1 is able to entirely complement the requirement for csp genes, strain MW_ΔCD-IF1^Ec (see above) was transformed with chromosomal DNA from B. subtilis csp double-deletion strains 64BC and 64BD and selected for deletion of all csp genes. Although this experiment has been repeated numerous times with appropriate control transformations, only a few if any spontaneous resistant transformants were obtained, and all still harbored at least one chromosomal csp gene. Transformation of strain MW_ΔCD-IF1^Ec with chromosomal DNA prepared from the chromosomal csp triple-deletion mutant B. subtilis 64BCDbt that carries a copy of cspB on the multicopy plasmid pDG148 in trans (designated pDGcspB [13]) resulted in transformants that indeed carried no chromosomal csp genes. However, PCR analysis of the pDG148 derivative plasmids present in all of these single colony transformants revealed that in some of these strains pMW_infA^Ec-1 had been entirely replaced by pDGcspB, while in others a mixture of both plasmids (which differ only in the infA or cspB insert and can be distinguished by using PCRMix3_P_spac-infA^Ec and PCRMix3_P_spac-cspB, respectively, as detailed in Table 2 and Fig. 3) was present. These results suggest that, under the conditions tested, a full complementation of CSP functionality by IF1 appears to be impossible in B. subtilis.

The experiments presented in this study show that growth and cell differentiation defects acquired by B. subtilis as a result of a combined deletion of the cold shock genes cspB and cspC (13, 47) can be cured by heterologous expression of the translation initiation factor IF1 from E. coli. These findings demonstrate that, under the conditions examined, IF1 is able to carry out function(s) that are usually mediated by CspB and/or CspC in B. subtilis and suggest a first potential connection between CSPs and the initiation of translation that is believed to represent the bottleneck of bacterial adaptation during cold shock (7, 3, 21, 6). Although the precise nature of this in vivo complementation remains to be elucidated in detail by future experiments, we believe that the results presented here have some interesting consequences for our understanding of CSP function(s) and the function(s) of IF1 in vivo, which are briefly discussed below.

CSPs as alternative translation initiation factors.

It is an accepted model that the initiation step in bacterial protein biosynthesis requires translation initiation factor IF1, which binds directly to the 30S subunit of the ribosome, as demonstrated recently by cocrystallization experiments (5; reference (14) and references therein). However, it has been shown that some predominantly gram-positive bacteria, such as, for example, Bacillus stearothermophilus, apparently do not contain an IF1 homolog (24). If this is correct, these species must have found a way to functionally replace IF1, a possibility that is not surprising in light of several differences between the gram-positive and gram-negative translational apparatus. Interestingly, all of these organisms appear to possess at least one copy of a gene encoding a member of the CSP family. In light of these data and the structural similarity of CSPs and IF1 as outlined in Fig. 1, the possibility that CSPs might act as alternative translation initiation factors in B. subtilis and related bacteria appears tempting. It has already been shown that in Streptomyces aureofaciens CSPs copurify with ribosomes (32). Moreover, we have recently discovered that B. subtilis CSPs colocalize with ribosomes in vivo, indicating that they function directly at or close to the ribosome (30, 47). We believe that these findings—the high structural similarity between CSPs and IF1, their apparently coincident subcellular localization, and the complementation results presented in this work—justifies further investigation of the hypothesis that CSPs might act as alternative translation initiation factors.

CSPs and/or IF1 as RNA chaperone(s).

However, there are alternative explanations for the results obtained here. Although rarely mentioned in the literature, there is an early work demonstrating that IF1, in addition to its specific binding to ribosomal decoding site during the translation initiation process, alters the structure of various oligonucleotides in terms of disrupting nucleic acid interactions in vitro (39). These data can be interpreted as a putative chaperone activity as shown for CspA from E. coli (20) and as suggested for B. subtilis CSPs (13). Furthermore, cross-linking studies have shown that IF1 binds to mRNA in vivo, although it is unclear whether this interaction occurs during the translation initiation step or reflects activities of nonribosomal associated IF1 comparable to RNA chaperones (41, 42). With these data given, we conclude that the in vivo complementation results presented in this work can be interpreted in accordance with previously published data supporting the RNA chaperone model for CSP function or, alternatively, give rise to a novel hypothesis of CSPs acting as translation initiation-like factors in B. subtilis and possibly other bacteria as well. Consequently, detailed studies to identify the in vivo interaction partners of CSPs in B. subtilis are under way.