Severely incapacitating mutations in patients with extreme short stature identify RNA-processing endoribonuclease RMRP as an essential cell growth regulator

Abstract

The growth of an individual is deeply influenced by the regulation of cell growth and division, both of which also contribute to a wide variety of pathological conditions, including cancer, diabetes, and inflammation. To identify a major regulator of human growth, we performed positional cloning in an autosomal recessive type of profound short stature, anauxetic dysplasia. Homozygosity mapping led to the identification of novel mutations in the RMRP gene, which was previously known to cause two milder types of short stature with susceptibility to cancer, cartilage hair hypoplasia, and metaphyseal dysplasia without hypotrichosis. We show that different RMRP gene mutations lead to decreased cell growth by impairing ribosomal assembly and by altering cyclin-dependent cell cycle regulation. Clinical heterogeneity is explained by a correlation between the level and type of functional impairment in vitro and the severity of short stature or predisposition to cancer. Whereas the cartilage hair hypoplasia founder mutation affects both pathways intermediately, anauxetic dysplasia mutations do not affect B-cyclin messenger RNA (mRNA) levels but do severely incapacitate ribosomal assembly via defective endonucleolytic cleavage. Anauxetic dysplasia mutations thus lead to poor processing of ribosomal RNA while allowing normal mRNA processing and, therefore, genetically separate the different functions of RNase MRP.

Figure 1

Clinical and x-ray characteristics of anauxetic dysplasia. A, far left and near left, Patient, aged 16.5, from family 2 (Horn et al. 2001), showing severe short stature with hyperlordosis (height of 74 cm). Upper right, X-ray of the left arm at age 7, showing extremely retarded carpal ossification corresponding to ∼3 mo bone age, as well as short and broad tubular bones. Lower right, X-ray of spine and pelvis at age 7, showing decreased vertical dimension of the ilia, unossified femoral necks, and small, irregular capital femoral epiphyses. B, X-rays of lower extremities at age 7 mo (top) and pelvis at age 2 (bottom) of patient 3, showing δ-shaped metaphyses with irregular borders and pelvic changes similar to those in panel A.

Figure 4

RMRP gene mutations identified in anauxetic dysplasia patients of families 1–3 and their evolutionary conservation in various orthologs. Affected members of family 1 were homozygous for a 14-bp insertion at position 111_112 (A), affected members of family 2 were compound heterozygous for the mutations +90_91AG→GC (B) and +14G→A (C), and patient 3 was compound heterozygous for the +90_91AG→GC mutation (B) and for a +254C→G mutation (D). E, Significant decrease in RMRP expression of the +14G→A mutant allele (*), as revealed by RT-PCR in the patient of family 2.

Figure 5

Scheme of the two-dimensional structure of RMRP and nme1 RNA and the results of nme1 mutant analysis in S. cerevisiae. A, Predicted two-dimensional structure of human RMRP RNA, showing the position of the anauxetic dysplasia mutations +14G→A and +254C→G. B, Predicted two-dimensional structure (Shadel et al. 2000) of yeast nme1 RNA, showing positions considered equivalent by sequence alignment (M1=+14G→A and M2=+254C→G) or structural alignment (M3=+14G→A and M4=+254C→G). C, Reduced colony growth and metabolic activity in M1–M4 colonies grown on YPD and YPG plates at 30°C and 34°C, as compared with wild-type colonies that show red color due to an ade2 marker. Because of a mutation in the ade2 gene, adenine biosynthesis is impaired, which leads to the accumulation of an oxidized (red) precursor in yeast with normal metabolic activity.

Figure 6

A–D, Assessment of RMRP expression and endonucleolytic cleavage for ribosomal assembly in native cells of one patient from family 2 (P) with anauxetic dysplasia and compound heterozygosity for the RMRP mutations +90_91AG→GC and +14G→A. E and F, Rescue of phenotype by wild-type hyperexpression. A, Expression of RMRP was decreased 13-fold in the patient's lymphocytes and threefold in his fibroblasts (B), compared with average levels in 11 healthy control individuals with (C) and without outliers (C′). Decreased endonucleolytic cleavage in the patient, demonstrated by the increased ratio of 5.8S rRNA bound to ITS-1 versus cleaved 5.8S rRNA of 1.44-fold in lymphocytes (C) and 1.8-fold in fibroblasts (D). E, Rescue of phenotype by wild-type hyperexpression, as shown by the dramatically increasing cell count up to 48 h after transfection (dark bars) in comparison to transfection with the empty vector (pale bars). F, Increase of cell count correlates with increasing endonucleolytic cleavage in ribosomal assembly as measured by the ratio of 5.8S rRNA bound to ITS-1 versus cleaved 5.8S rRNA.

Figure 7

Comparison of different RMRP mutants overexpressed in transiently transfected human wild-type fibroblasts, with reference to relative cell counts and endonucleolytic cleavage activity in ribosomal assembly and cyclin levels. A, Relative cell count normalized to the level at 12 h after transfection in anauxetic dysplasia mutants and the CHH founder mutation +70A→G, as compared with the wild-type transfected cells. B, Impaired growth correlating with decreased endonucleolytic cleavage activity investigated via the ratio of 5.8s rRNA bound to ITS-1 versus cleaved 5.8s rRNA normalized to 12 h after transfection. C, Decreased CCNA2 expression in naturally expressed anauxetic dysplasia mutations and less severe in the CHH founder mutation. D, Significantly increased CCNB2 expression in the +70A→G mutation. Note that the +14G→A mutation is not expressed in vivo.

Figure 8

Structural properties of a GNRA tetraloop and implications for RNA-protein interaction, as evident from the nutboxB-RNA structure in complex with the N-protein (Schärpf et al. 2000). The RNA is shown in stick presentation and the bound protein is schematically shown in red. Key nucleotides of the GNRA tetraloop are colored in green and cyan, and the respective sequence positions in RMRP are given in parentheses. The G and the A of the tetraloop form a sheared base pair which will be disrupted by a G→C mutation in RMRP. The role of the tetraloop structure for RNA-protein interactions is emphasized by the stacking of a tryptophan in the nutboxB-RNA complex with the N-protein.

Figure 9

Simplified overview of the cell cycle–related pathways impaired in anauxetic dysplasia and CHH. Right, The hereby affected RMRP gene encodes the untranslated RNA subunit of the ribonucleoprotein endoribonuclease, RNase MRP, which is essential for cell growth and division in yeast and, as our data suggest, also in humans. One function of the RNase MRP complex is the processing of the precursor of 5.8S rRNA, which is a subunit of the 60S ribosomal particle. Therefore, severe and moderate disruption (red flashes) of RNase MRP function in anauxetic dysplasia and CHH, respectively, impacts late-60S ribosomal assembly, resulting in a reduced capacity to synthesize proteins. As a secondary effect, cyclin A2, which promotes G1/S and G2/M phase transitions, is diminished, correlating with the magnitude of delay in the cell cycle. Left, The second function of RNase MRP complex in yeast and, apparently, in humans is the degradation of cyclin B2 mRNA, which is important for the exit of mitosis. In contrast to anauxetic dyplasia, the latter pathway is also impacted in CHH, as shown by increased cyclin B2 mRNA levels. Since cyclin B2 overexpression from different mechanisms contributes through alterations of the spindle checkpoint to the chromosomal instability observed in some cancers, our findings could also explain why only CHH, and not anauxetic dysplasia, is associated with proliferative bone marrow dysfunction and susceptibility to cancer.

Figure 2

Scheme of rRNA processing in yeast and humans. A, Deletions of nme1, the RMRP RNA gene homolog in yeast, impacts late-60S ribosomal assembly via defective endonuclease cleavage of the precursor subunit 5.8S rRNA at the ITS-1 A3 site (red arrow). The eukaryotic pre-60S particles undergo maturation via several endonucleolytic cleavages, resulting in 25S, 5.8S, and 5S rRNA components in yeast and 28S, 5.8S, and 5S rRNA components in humans. In yeast, the 5.8S rRNA consists of two main species, 5.8S_L and 5.8S_S, with a ratio of 1:10 long to short, generated by two different pathways. The nme1 gene product is required for the production of the 5.8S_S species; hence, depletion leads to increased levels of 5.8S_L and reduced levels of 5.8S_S. B, Assumed cleavage site of human RNase MRP (red arrow) and location of Taqman probes used to measure endonucleatic cleavage in mutants.

Figure 3

Summary of the genome scan for anauxetic dysplasia or spondylometaphyseal dysplasia (SMED) with use of GENEHUNTER-PLUS (Kong and Cox 1997). Upper panel, Multipoint parametric LOD-score analysis across all autosomes, from pter to qter, with use of a fully penetrant autosomal recessive model. For clarity, only LOD scores >−2 are shown. Maximum LOD score was 3.012 at markers D9S171 and D9S1874 on chromosome 9p. Lower panel, Nonparametric multipoint Z score. Again, the maximum Z-score was obtained for markers in the same region on chromosome 9p. Dotted lines indicate chromosome boundaries.